;!c,^::r:x.;:;;|;;Vi!!v fc::fi^9{X^:;^;';fSjC>5f^lf'';'^''''t;'"!:r!: Museum of Comparative Zoology HERPETOLOGY LIBRARY Ernst fJlayr Library ^ ^^^J^ ^^ ^■fr-m?>aratve Zoology /J p-' ^-^^ — A HISTORY OF BRITISH FOSSIL REPTILES SIR RICHARD OWEN, Iv.C.B., E.R.S., Etc., FOREIGN ASSOCIATE OF THE INSTITUTE OF FEAKCE (ACADEMY OF SCIENCES). VOL. III. LONDON : CASSBLL & COMPANY LIMITED, LA BELLE SAUVAGE TAED. 1849—84. PEIMTED BT J. E. ADtARI), BAETHOLOMEW CtOSE. CONTENTS. PAGE-INDEX. FOSSIL REPTILIA OY THE LIASSIC FORMATIONS. CHAPTER I. Order— SAUROPTERYGTA. § 1. Genus— Pi. EsiosAURUs. Species — Plesiosaiirus doUchodeirus . — homalospondyhis — rostratiis — rugosus 1 12 20 34 CHAPTER II. Order— ICHTHYOPTERYGIA. Genus — Ichtiiyosauhus . 41 A. Introduction 41 B. Osteology 45 Species — Ichthyosaurus breviceps 67 — communis 68 — intermedius 70 — platyodon 73 — lonchiodoH 75 — longifrons 76 — latifrons 76 — acutirostris 78 — teniiinostris . 7Q — longirostris 82 — latimanvs 83 — brachyspondylus 85 Conclusion . 85 IV PAGE-INDEX. CHAPTER III. Order— DINOSAURIA. Genus — Scelidosauuis Species — Scelidosaurus Harrisunii PAGE 89 92 (Supplement No. II.) MeZOZOIC L.\CE11TILIA Geuus — EcuiNODON Species — Echinodon becclesii 126 12G 126 CHAPTER IV. Ordkr— CROCODILIA. Fainilv— Protosuciiii 129 Geuus — Teleosauuus 130 Species — Teleusiiiiriis C/ta/jinauni 130 — cudoineiisis 139 — brevior 140 — tatifrons 141 — astheaodeirus 141 Geuus — Steneosaukus 143 Species — Sleiieosuurus Geoffroyi 144 — luticeps 145 — temporalis 145 Geuus — Plesiosuouus 146 Species — Plesiosuchus Mansellii 146 CHAPTER V. Order— SAUROPTERYGIA. Genus — Pliosaukus .... . 152 Species — FUosuurus yiandis . 164 brac/ii/deirus . 157 — truchuiUerius . 158 — portlandicus . 163 PAGE-IiNDEX. CHAPTER VI. Order— DINOSAURIA. Genus — Megalosaurus Species — Megalosaurus Bucklandi GeuUS — BOTHIIIOSPONDYLUS Species — Bothriospondylus magnus PAGE 166 166 172 , 172 ICHTHYOPTEEYGIA. Ichthyosaurus fortimanus — longimanus (Supplement.) 176 176 Sauuopterygia. Plesiosaurus macrocephalus — brachycephalus (Supplement.) 177 178 (Supplement.) Cheloxia. Genus — Pleurosteenon . Species — Pleurosternon concinnum — emarginatum — ovatum — latisculatum Genus — Platemys. Species— PI ateiiiys Mantelli Dixoni Genus— Chelone. Species^CAe/one costata ■ — gigas 179 179 183 184 185 186 187 187 188 Order— LABYRINTHODONTIA. Genus — Labyrinthodon . Species — Labyrinthodon Jaegeri — leptognatkus — pachygnathua — scutulatus ■ Footprints of Labyrinthodon 189 190 191 192 195 197 CONTENTS. SYSTEM-INDEX. Order— CHELONIA. Genus — Pleueosternon — Platemys — Chelone PAGE 179 185 187 Order— LACERTILIA. Genus — Echinodon. Species — Echinodon hecclesii 126 Order— CROCODILIA. Family- PRGTOSUCHII. • •.... . 129 Genus — Teleosaurds. Species — Teleosauru^ ■ Chapmnnni ..... . 130 — cariomensis ..... . 139 — brevior ..... . 140 — latifrons ..... 141 — asthenodeirvs ..... 141 Genus — Steneosauehs ...... 143 Species — Steneosaurus Geoffroyi ..... 144 — laticeps ..... 145 — temporalis ..... 145 Genus— Plesiosuchus. Species — Plesiosuchus Mansellii ..... 146 Order— DINOSAURIA. Genus — Scelidosadkus Species — Scelidosaurus Harrisonii Genus — Megalosaueus Species — Megalosaurus Bucklandi Genus — Bothriospondylus Species — Bothriospondylus magnus 89 92 166 166 172 172 SYSTEM-INDEX. vn Order— SAUROPTERYGIA. Genus — Pliosaurus Species — Pliosaurus grandis — brachydeirus ■ — trochanterius — portlandicus Genus — Flesiosaukus Species — Plesiosaurus dolichodeirus — homalospondylus — rugosMS PAGE 152 154 157 158 163 1 1 12 34 Order— ICHTHYOPTERYGl A. Genus — Ichthyosaurus Species — Ichtliyosaurus breviceps communis intermedius platyodon loncliiudoii longifrons latifrons acutirostris tenuirostris longirostris latimanus bracliyspondylus fortimanus longimanus . 41 67 68 70 73 75 76 76 78 79 82 85 85 176 176 Order— LABYRINTHODONTIA. Genus — Labyrinthodon .... . 190 Species — Labyrinthodon Jaegeri . 190 — leptognathus . 191 — pachynathus . 192 — scutulatus . . 195 Footprints of Labyrinthodon . 197 A HISTORY BRITISH FOSSIL REPTILES. THE FOSSIL REPTILIA OF THE LIASSIC FORMATIONS. CHAPTER I. Order— SAUROPTERYGIA, Oim/. Genus — Plesiosaurus, Conybeare. Species — Plesiosaurus doIicJiodeirus, Conybeare. (Tabs. I— IV.) Of the Plesiosaurus dolichodeirus, Conyb., the first described and the typical species of the genus, three more or less entire specimens have come under my obser- vation, which have been obtained from the Lower Lias of Lyme Regis and Charmouth, Dorsetshire. One of these, formerly in the possession of the late Duke of Buckingham and now in the British Museum, was the subject of Conybeare's original description.* A second, in the British Museum, is figured by Buckland in his ' Bridgewater Treatise,' vol. ii, pi. xix, fig. 2 ; the third, also in the British Museum, is the one which I have selected for illustration in the present Chapter (Tabs. I and II). In this the vertebral series is entire ; there is no break in the long cervical region, as in the other two specimens ; its perfection, in this respect, satisfactorily shows that the head is at, or nearly at, the correct distance from the trunk, with the neck outstretched, in the two former specimens, the greater completeness of which, in regard to the Hmbs, supplies what is wanting in this respect in the present skeleton (see Tab. I, figs. 2 and 3). The condition of the vertebral column in the originally described or type-specimen of the Plesiosaurus doUcJiodeirus is such as to suggest that the carcass, after it sank to • 'Transactions of the Geological Society,' 2nd series, vol. i, p. 381, pi. xlviii. b 2 BRITISH FOSSIL REPTILES. the bottom, had been preyed upon by some contemporary carnivorous marine animals. It seems as if a bite of the neck had pulled out of place the eighth to the twelfth vertebrae. Those at the base of the neck have been scattered and displaced, as if through more " rugging and riving." Some creature which has had a grip of the spine, near the middle of the back, has pulled to one side all the succeeding vertebrae of the pelvis; their adhesion to that part and, more or less, to each other, being retained. This wrench would expose the abdominal viscera, a tergo, where we now see the upper or inner surface of the abdominal ribs or sterno-costal arches. The intermediate and succeeding portions of the vertebral column retain their natural relative positions, as in the prone position of the carcass ; and the skull, scapular arch and appendages, pelvic arch and appendages, and the tail, show respectively their relative positions as in the entire animal. Many of the otherwise undisturbed vertebrae, however, have turned, so as to present their most extensive surface to the direction of the slow, cosmical, compressing force operating on their imbedding stratum. This is the case with the first twenty cervical vertebrae in the specimen Tab. I, which appears to have settled in the Liassic mud back downwards, their spines being turned toward the right side ; beyond the twenty-first cervical the vertebrae have rotated in the opposite direction, presenting more or less of a side view, with the neural arch and spine turned to the left ; but most of the spinous processes have been removed with the matrix in the original exposure of the specimen. The trunk preserves the supine position, exposing the broad coracoids (52), and pubes (G4), with scattered, intervening, abdominal ribs. Part of the left pectoral fin (53—56) is in situ; a smaller part of the corresponding pelvic fin (65—67) lies across the pelvis. No partial force has operated after interment to dislocate any of the vertebrae, save the few terminal ones of the tail, which have disappeared, probably dragged away with whatever tegumentary expansion may have there represented a caudal fin. In the specimen figured by Dr. Buckland * the skeleton, as it is exposed to view, lies prone ; the vertebra-, whilst their matrix was in the state allowing them to turn, have presented their largest surface to the direction of superincumbent pressure, the spines of those at the basal half of the neck being turned down or toward the right side, while those of the dorsal vertebrae have yielded in the opposite direction, both kinds presenting more or less of a side view. The thoracic ribs have slipped some way from their articulations, yet preserve, in the main, their relative positions, in serial succession. The anterior dorsals overlie the coracoids, and the posterior dorsal and sacral vertebrae overlie the dislocated parts of the pelvis. One of the thickened, short, and straight sacral ribs abuts against the right ilium. Upwards of thirty caudal vertebrae extend, in nearly a straight line, from the sacrum. The vertebrae at the fore part of the neck have been displaced, and in great part lost. Of the head little is visible, save the mandibular rami. The bones of both fore and hind paddles on the * Op. cit., vol. ii, plate x, fig. 2. LIASSIC PLESIOSAURS. 3 right side are in an instructive state of preservation, especially those of the hind fin, which exemplifies the slight superiority of length as compared with the fore fin, cha- racteristic of the present species. The following are admeasurements of corresponding parts of the three skeletons above mentioned, that of " entire length" being now capable of being given by reason of the integrity of the cervical region of the spine between the head and the pectoral or scapulo-coracoid arch. Conybeare's Bnckland's Specimen specimen. specimen. Tah. I. Ft. ill. ines. Ft. in. lines. Ft. in. lines Entire length (two or three inches of the tail wanting ?) 9 5 0 9 8 0 8 9 0 Length of head . 0 8 6* 0 8 2 0 8 6t )> neck . 4 0 0 4 10 0| 4 10 0 j» trunk, from fo re part of sternum to end of ischium . 3 0 0 3 6 0 2 8 0 >) tail . . 2 0 0 2 4 0 (imperfect.) >) pectoral limb I 10 0 1 9 0 (imperfect.) )j pelvic limb 2 0 0 2 2 0 (imperfect.) >» humerus . . 0 7 0 0 7 6 0 7 6 »j femur 0 7 0 0 7 4 (wanting.) ») •radius 0 3 0 0 3 0 0 3 6 )) tibia . 0 3 0 0 3 3 >> mauus 1 0 0 1 0 o§ jt pes . 1 2 0 1 3 0 Breadth of pubis (transverse diameter) 0 .') 0 (obscured) 0 6 0 Length of ischium (longitudinal diameter) . 0 4 6 (obscured) 0 5 0 From the amount of concordance in the dimensions of the above three skeletons, it may be inferred that the average length of the mature animal of the present species of Plesiosaurus was between nine and ten feet. Two specimens of certain portions of the skeleton, now in the British Museum, one of which is the subject of Tab. Ill, support this conclusion. Vertebral characters (Tabs. Ill and IV). Cervical series. — The cervical vertebrae of the Plesiosaurus dolichodeirus, at the fore part of the neck (Tab. Ill, fig. 1, a, c4— 7), have the centrum (c) of a length equalling the breadth of the articular end ; but the dimension of breadth increases in a greater * Estimated at that of the lower jaw. ■\ Nine inches, if the angle of the jaw he restored, as dotted in Tab. IL X It may be that the head has been drawn a little forward in the displacement of the anterior cervical vertebrae. § Some of the terminal phalanges are here estimated for. 4 BRITISH FOSSIL REPTILES. ratio as the vertebrae enlarge and recede in position; so that the length of the centrum may be one sixth less than the breadth at the middle (Tab. Ill, figs. 4, 5, 6) and hind part of the neck. There are, however, varieties in this respect, and the equality of length to breadth of centrum is maintained through a greater extent of the neck in some specimens than in others. The vertical diameter of the middle of the terminal articular surface (ib., fig. 6 c) is less by nearly one fourth than the breadth of the same. The sides of the centrum are longitudinally concave (ib.,fig. 5), as is also the under part (ib., fig. 4), but in a minor degree in the middle and posterior than in the small an- terior cervicals (ib., fig. 1, e). The costal surface is at the lower part of the side of the centrum ; it is narrow vertically, in proportion to its length in the anterior cervicals, but gains in vertical extent without being elongated in the same degree, and consequently occupies a larger corresponding extent in the middle cervicals (ib., fig. 4, ;j/),* in which tlie costal surface exceeds one half the length of the centrum ; it is divided by a longitudinal cleft, and is situated a little nearer the posterior than the anterior surface of the centrum. In the third to the seventh cervicals (ib., fig. 1, c7) the costal surface is separated by a tract exceeding its own vertical diameter from the neurapophysial surface (ib., fig. 1,«); in the succeeding cervicals the intervening tract equals the costal surface (ib., fig. 4, c ) ; and the interval is never less, and is sometimes more, in the cervical vertebra? to near the base of the neck. The terminal articular surface (ib., fig. 6) is moderately convex at its periphery and very gently concave in the rest of its extent, with a small central, often transversely linear, impression in the centre. f The free surface of the centrum is finely rugose in the smaller anterior cervicals, and is not smooth in any of the others ; towards the articular ends the roughness is more marked, by irregular narrow risings and groovings, which become more longitu- dinal in direction in the succeeding cervicals (ib., figs. 4, 5). The under surface (fig. 5) is concave transversely from the costal pit (pi) to the two venous openings, and is convex between those openings, which divide the surface pretty equally into three parts, Lengthwise, as already stated, the under surface is gently concave. The neurapo- physial surface is less angular than in some other species, the lower angle being rounded off, making the lower border approach to a curve. Anchylosis of the neural arch with the centrum seems to have been complete in the anterior cervicals of the specimen figured in Tab. Ill, fig. 1 ; and I have not yet seen a cervical centrum of the present species from which the neural arch had become detached, save by fracture. The zygapophyses are proportionally large ; the anterior ones (ib., c, z ) extend forward, in the anterior vertebrte almost immediately above the centrum, overhanging the * Here obscured by the confluent base of the rib. f " The concavity again slightly swelling in a contrasted curve near the middle of the circular area," (Couybeare's first Memoir, p. 582, April, 1821) is the character of the terminal articular surface in the riesiosaurus arcuatvs, from the Lias in the neighbourhood of Bristol. LIASSIC PLESIOSAURS. 5 posterior two fifths of the centrum in advance ; the articular surface, the length of which equals two thirds that of the length of the centrum, looks obliquely upward and inward ; that of the posterior zygapophysis has the reverse aspect. As the cervicals approach the back the zygapophyses diminish in relative size, and their articular sur- faces become less horizontal. The posterior zygapophysis (Tab. Ill, fig. 4, y) over- hangs a small part of the end of the succeeding centrum, and the neurapophysis (ib., fig. 4, n ) rises with a deeper concavity at the back than at the fore part. That this anchylosis had not occurred in the similarly sized and in the larger specimens of the cervical vertebrae of the Flesiosaurus described and figured by Conybeare in his first famous Memoir* is due to their having been derived from a younger specimen of a larger species from the Bristol Lias, probably Plesiosaurus arcuafus. The neural spine (Tab. I, ns; Tab. Ill, fig. A,ns) arises narrow between the bases or back part of the prezygapophyses (^), and its base extends, increasing in thickness gradually to near the back part of the postzygapophyses (j')- The height of the spine averages half the vertical extent of the entire vertebra from its summit to the lower level of the centrum, being rather shorter in the anterior cervicals and exceeding that length at the base of the neck. In the anterior cervicals the contour of the neural spine extends from the fore part of the base, in a curve increasing in convexity at the upper part, and terminating by a rounded apex overhanging, in the foremost vertebrae (as at c4, Tab. Ill), the concave contour of the hinder border. The upper part of the spine becomes more squared as the spine itself gains in height, in the larger posterior cervicals, by the increasing fore-and-aft extent of their upper part, as in fig. 4, ns. The pleurapophysis of the axis (Tab. Ill, fig. 1, xpl) has its posterior angle extended backward ; that of the third cervical has its anterior angle also produced forward, but in a minor degree. Both angles continue to be more produced in the succeeding vertebrse, but the front one most so, until, in the fifth cervical, they are equal in length ; the hinder one then elongates, but they do not touch or overlap the contiguous pleurapophyses until about the tenth cervical vertebra. The extent of this terminally dilated or extended border of the riblet exceeds that of the diameter of the same from its upper articulation outward or downward. The line of- articulation is discernible in most of the anterior vertebrae, but in fig. 1 coalescence has com- menced, if it be not complete, as in figs. 4, 5, 6, pi in which the expanded part of the pleurapophysis has been broken off, showing the approximated head and tubercle adapted respectively to par- and di-apophysial divisions of the costal surface. In * " Notice of the discovery of a new fossil animal, forming a link between the Ichthyosaurus and Crocodile, together with general remarks on the osteology of the Ichthyosaurus; from the observations of H. T. de la Beche, Esq. F.R.S., M.G.S., and the Rev. W. D. Conybeare, F.R.S., M.G.S. (Read April 6th, 1821.) Drawn up and communicated by the latter." The observations on the vertebral characters of the new reptile are said to have been made " on the organic remains contained in the Lias in the vicinity of Bristol" (p. 559). 'Transactions of the Geological Society of London,' first series, vol. v. 6 BRITISH FOSSIL REPTILES. some vertebrae a low and narrow ridge extends from the neur- to the pleur-apophysial surfaces, as at c, fig. 4. The degree of concavity of the sides of the centrum in the anterior cervicals, exposed in the specimen figured in Tab. Ill, fig. 1, has been exaggerated by the pressure to which it has been subject, the effects of which are more conspicuous upon the skull : the cancellous mid-part of the centrum has opposed less resistance than the compact articular ends. The atlas (a) has been disarticulated from the occipital condyle (l) ; the hemispheric articular cup is thus well displayed, with its smooth and shining surface. The coales- cence of the centrums of the atlas [a] and axis (a) is complete. A tubercle from the side of the centrum of the atlas represents the pleurapophysis ; its neural arch is broken away ; that of the axis developes a spine similar to and but little smaller than that of the third cervical. The dimensions of the seven vertebrae here preserved in connection with the skull will be seen in Tab. Ill, fig. 1, where they are figured of the natural size. The average dimensions of a cervical centrum of the present species, from the middle and basal half of the neck, are, as in figs. 4, 5, 6 — In. lines. Length . . ; . 14 Breadth of articular surface 16 Height of middle of ditto ......... 13 Length of costal pit ..... . ... 0 6 Transverse diameter of outlet of neural canal ..... 0 6 Dorsal series. — The transition from the cervical (Tab. I, c) to the dorsal series (ib., d) is effected by the usual elevation of the costal surface by gradational steps, continued through about five vertebree, until a single costal surface is presented by a large diapophysis from the neural arch. The number of cervical vertebrae so defined in the specimen figured in Tab. I is forty-one. In the first dorsal, characterised by the dia- pophysial support of the rib (Tab. IV, figs. I and 2, d), the non-articular part of the centrum is smoother than in the cervical vertebrae, the ridges or rugae occupying a smaller extent near the two ends, where they indicate the attachments of the capsular ligaments. The longitudinal concavity between the two ends is uniform and rather more than in the cervicals. The venous foramina are wider apart and not divided by any special transverse convexity on the under surface of the centrum. A vertical ridge leads from the side of the centrum (ib., c) to the under part of the diapophysis (ib., «/), nearer the hind than the fore end of the centrum. The diapophysis is convex and longest superiorly ; the fore part is rather hollowed, the hind part flattened, and both converge to the ridge forming the shorter under surface. The articular surface C^) of an irregular oval form, with the small end down- LIASSIC PLESIOSAURS. 7 ward, looks downward, outward, and a little backward, the process being slightly inclined that way. The margin of the articular end of the centrum is better defined than in the neck; about a line's breadth is, as it were, shaved off; the rest of the sur- face (fig. 2, c) is very slightly concave, sometimes undulated, always nearly flat, and with a small central depression, or a tendency there to a tubercle. The length of tlie dorsal region in the skeleton (Tab. I) 8 feet 9 inches long, is 2 feet G inches ; the number of dorsal vertebrae is twenty-one. Sacrum. — Two vertebrae (ib., s) succeeding the dorsals are distinguishable, through the greater thickness and straightness of their short pleurapophyses, as sacral ; these elements abut against, or afford ligamentous union to, the iliac bones. Caudal series. — The caudal vertebrae (Tab. I, c d, Tab. IV, figs. 3 — 9) are shorter in proportion to their breadth than the others ; the centrum approaches to a cubical figure, the under surface (figs. 4 and 7) becoming broad and flattened ; and the contour of the terminal articular surfaces shows a similar tendency to flattening, giving a transversely extended quadrate figure, with the angles rounded off (figs. 5 and 8) ; the margin is thicker, more rounded off, less defined than in the dorsal vertebrae. The articular surface itself is more concave than in the antecedent regions of the backbone, and becomes deeper in the terminal subcompressed vertebrae (fig. 9) ; the movements of the tail in swimming having been helped here by a greater amount of yielding intervertebral substance, approaching in the same degree to the condition of the spine in fishes. The costal surface (figs. 3, 6, pi) is elliptic, with the long axis subvertical, the margin prominent, the cavity simple and rough for the ligamentous attachment of the ril:)let ; it is situated on the upper half of the centrum close to the neurapophysis, the outer end of the base of which contributes to the upper part of -the margin in the anterior caudals (fig. 5, d). The pleurapophyses in this region (Tab. IV, 5, pi) do not expand terminally, as in the neck ; they are short, thick, and straight, simulating transverse processes ; their non-confluence with the centrum exemplifies the minor vigour of vital co-ossifying influences in terminal parts. The heemapopliysial surfaces (fig. 3 A') impress the inferior angles of the posterior surface of the centrum ; occasionally, where a haemapophysis has become anchylosed and broken off, its adherent base gives the appearance of a process from that part.;t)f the centrum (ib., figs. 6, 7, S, /*')• The venous foramina are at the lower part of the sides of the centrum. The neural arch (figs. 3, 5, «) rapidly diminishes in size and in the length of the neural spine, ns. The zygapophysial surfaces become more vertical, the anterior, z, looking inward ; the posterior zygapophyses, z', are the first to disappear. Tlie haemapophyses (fig. 5, h) are free, and were ligamentously connected with the centrum above and with each other below, circumscribing there the haemal canal. The proximal surface is expanded, with a subtriangular facet cut obliquely at the anterior part for articulation with its own surface, and with a smaller, less definite 8 BRITISH FOSSIL REPTILES. surface posteriorly for the intervertebral substance and a small part of the succeeding centrum, where a slight expansion of the everted border of the articular surface is the sole indication of such hsemapophysial junction. In the terminal vertebrae these surfaces with the hsemapophyses have disappeared, and the centrum, now showing a compressed form, supports only a contracted, anchylosed, seemingly exogenous neural arch, which finally disappears. The following are transverse diameters of the centrum in different regions of the spine, in the specimen, 8 feet 9 inches long, of Plesiosaurus doUchodeirus, figured in Tab. I : In. lines. Tenth cervical vertebra ....... 1 0 Middle dorsal ditto 2 0 Tenth caudal ditto 13 Cranial characters (Tabs. II and III). The skull in this skeleton presents, what is rare, the side or profile view (Tab. II) like that of the succeeding anterior cervicals. Its upper part is much injured. The following bones are recognisable : — mastoid 8, tympanic 28, squamosal 27, malar 26, maxillary 21, premaxillary 22; the end of the long pterygoid is seen at 24, abutting against the lower end of the tympanic. But little of the composition of the mandible is discernible ; the tightly closed jaws show the extent of the interlocking of the long, slender, curved, and sharp-pointed teeth. Of those of the lower jaw the crowns of upwards of twenty may be traced ; the longest occupying the middle three fourths of the series, and the largest of these being the foremost. In some parts of the series two teeth pass into the same dental inter- space of the opposite jaw. The admirably wrought-out specimen figured in Tab. Ill, fig. 1, exhibits the upper surface of the somewhat crushed skull. Of the basi-occipital a part of the upper surface (l) and of the single median convex condyle is shown. The exoccipitals (ib., 2) preserve their connection with the lateral and upper parts of the basi-occipital, and show the . surfaces — seemingly sutural- — from which the superoccipital (ib., 3) has been displaced. These surfaces (2, 2) are thick' and triangular; they are parallel with the midiMe of the foramen magnum, the lower half of which is formed by the basi- and ex-occipitals. From the outer and back part of the exoccipital the paroccipital process (4, 4) is continued ; of subtriedral form, long, slender, and tapering to a thin I'ounded apex : the outer side appears to be sutural, and that of the left side is applied to the tympanic (ib., 28) : the length of this process is 8 lines. The breadth of the occiput, outside of the exoccipitals, is 1 inch .'j lines ; that of the foramen magnum is 6 lines. That the rough triangular upper surfaces of the exoccipitals are natural, not the result of frac- LIASSIC PLESIOSAURS. 9 ture, I infer from their being on the same level, and from the corresponding surfaces being presented by the single arched bone (3) representing the superoccipital. This has been displaced by the pressure operating not quite vertically, but with an incli- nation from the left to the right, which has turned the spines of the cervical vertebrae to the right, and which has so far displaced the superoccipital in the same direction that it lies with its concavity or arch embracing, as it were, the right superoccipital, this concavity having formed the upper half of the foramen magnum. The apex of the superoccipital lies beneath the right branch of tlie parietal ; the outer surface of the piers of the superoccipital arch is moderately smooth and convex ; the breadth of the base of the arch is 1 inch 3 lines, that of the span of the arch is 5 lines. The parietal (ib., 7) is thick and transversely extended posteriorly, where it is overlapped by the mastoids (s), anterior to which it contracts to form the crest be- tween the temporal fossae. The crest is interrupted by the parietal foramen (/), an- terior to which it is resumed for a short extent — 3 lines, before the frontal suture. The total length of the parietal is 1 inch 1 1 lines ; the length of the crest is 1 inch 3 lines. Tlie thick and rather rugged hinder bifurcate part of the parietal is overlapped or embraced by the mastoid (s), and these bones curve outward and backward to articulate with the squamosal (,27) find with the tympanic (28), which is continued in the same direction to the joint of the mandible (29). AH these bones together form a strong arch, curved backward in the present specimen, but owing its horizontal position to the posthumous pressure, and having the piers of the arch directed down- ward as well as outward and backward in the natural state. The sutui'e between the frontals (11) remains, and that between the postfrontals (12) and the expansions of the parietal {'') upon the sides of the cranium may be traced. There is a smooth superorbital (11') between the rougher frontal and the orbit, unless the fissure defining them be a fracture and not a suture. The external facial plate of the prefrontals is rough ; it overlaps the fore part of the frontal and part of the nasal, and extends to the small external nostril. The nasals (15) overlap the fore part of the frontals, and extend about as far in advance of the nostrils as they do behind them, continuing the median ridge from the frontals forward, in which, however, the median suture is visible. The outer surface of the maxillarics and premaxillaries shows a kind of granular rugosity, which subsides in the maxillary as this bone (21) extends beneatli the orbit. The limits of the lacrymal (73) are not definable. The malar (26) forms the hinder half of the suborbital boundary; its surface is smooth, and increases in breadth to beyond the orbit, when it contracts and becomes rugous where it joins the postfrontal (12) and squamosal (27). The bony boundary between the orbital and temporal cavities is crushed and much cracked : but the outer end of a postfrontal or postorbital is wedged into the squamous union of the malar and squamosal. The latter (27), of a tri-radiate form, curves from the malar round the outer and back angle of the temporal fossa, and extends backward upon the tympanic : the ray directed c 10 BRITISH FOSSIL REPTILES. mesiad, and overlapping the mastoid (s) and tympanic (28), is the longest, and termi- nates in a point : the surface of the bone is smooth. The temporal fossae are broader than they are long. At their forepart the parietal side-wall of the cranium expands as it advances, and is continued into the postfrontal or postorbital partition. The orbits are rounded anteriorly, and both the upper and under parts of the frame make an angular junction with the straight hinder pai't. The nostrils have the usual small size and backward position. In both orbits some of the thin sclerotic plates of the eyeball («, s) are preserved ; this is the first specimen in which I have had evidence of this structure. The interlocking of the teeth of the upper and lower jaws, through the singular care and skill devoted by Mr. Harrison to the removal of the matrix, is peculiarly well displayed in this instructive fossil. The foremost tooth in each premaxillary make a pair, which curve forward and downward between the two foremost teeth of the lower jaw, the premaxillary teeth slightly diverging as they descend, Tab. Ill, fig. 3. The succeeding premaxillary teeth, four in number, alternate with mandibular ones. I cannot make out with certainty the maxillo-premaxillary suture, but the fifth tooth, counting backward, seems to be near to or upon it. The second premaxillary tooth is double the size of the first ; the third, fourth, and fifth gradually diminish ; the sixth (first maxillary ?) is small ; the seventh tooth suddenly resumes- the size of the second ; and the eighth, of nearly equal size, curves down close to the seventh, and the two are interposed between the inter- space of opposite mandibular teeth. From four to five smaller teeth are traceable behind the eighth, and there may have been more in the upper jaw. Of the lower jaw ten teeth are shown on each side ; the second, third, fourth, and fifth are the longest and largest, as in Tab. II. In general, the teeth of the upper jaw are separated by intervals allowing the passage of those of the lower ; the teeth of the foremost premaxillary pair being closer together ; and those of the foremost mandi- bular pair being wider apart. They all present the usual generic character of crown — long, slender, curved, pointed, circular in transverse section, with the enamel finely but definitely ridged longitudinally. The longest exserted crown measures ten lines, the shortest four lines, the thickness being in proportion. The true number of the teeth in the lower jaw is yielded by the specimen of the dentary bone. Tab. Ill, fig. 2, in which twenty-five alveoli are shown on one side, and twenty-four on the other. The size of the alveoli, and the extent of their interspaces, are greatest at the anterior half of dentary. The small successional teeth at the posterior part of the series are so advanced as to look like a double row at that part. A longitudinal groove or depression at the inner side of the base of the alveoli lodged the thicker mass of the vascular gum overlying the matrices of the successional teeth. The skull of the Plesiosaurus dolichodeinis is broad in proportion to its length, with LIASSIG PLESIOSAURS. 11 a broad and short muzzle, of an equilateral triangular figure if the transverse lines across the fore part of the orbits be taken as the base, the two sides converging to the rounded apex in almost straight lines, with a feeble indication of a constriction where the maxillo-premaxillary suture seems to be. The contour is undulated by the expan- sions for the sockets of the larger teeth, which produce risings, with intervening furrows on the granulate alveolar borders of the jaws. The mandibular rami converge to their terminations at the symphysis, which is not prolonged or expanded. The specimen (Tab. Ill, fig. 1) from the Lias of Charmouth was wrought out of its matrix by the estimable discoverer of the Liassic Dinosaur [Scelidosaurus) described in a former Monograph. It is an admirable example of patience, pains, and skill ; in the bestowal of which, for the furtherance of science, upon the fossils roughly wrought out of the quarries in his neighbourhood, Mr. Harrison found solace during the long and trying illness which confined him to his bed, until his final release by death. Pectoral and pelvic limbs (Tab. I, figs. 2, 3). To complete the characters of Plesiosaurus dolichodeiriis I have reproduced, in out- line, the bones of the pectoral, fig. 2, and pelvic, fig. 3, limbs, as they are preserved in the type-specimen. The humerus, % shows rather more convexity at tlie anterior border, and a deeper concavity at the posterior border than in some other species {Plesiosaurus Hawhinsii, PL luacrocejjJialus,^ PI. riigosus, e. g.). The radius, 54, and ulna, 55, are of equal length ; the ulna not being shorter than the radius, as in PL Hawlcinsii (Tab. XIV, fig. 6) ; the ulna has not the olecranal process or epiphysis, as in the PL rurjosus (Tab. XIV, fig. 2, 55') ; and both antibrachial bones are less broad^ in proportion to their length, than in the PI. macrocephali(s-\ (Tab. XIV, fig. 4). The carpus shows seven bones, four in the proximal, three in the distal row ; their homo- logies will be pointed out in the description of Plesiosaurus rur/osus. The metacarpal of the first digit (Tab. I, fig. 2, i), answering to "pohex,'' supports at least three phalanges ; that of the fifth digit, seven phalanges ; the metacarpal of each of the others, six phalanges, but the terminal ones may be wanting in some. The pelvic fin (fig. 3) is rather longer than the pectoral one ; in both fins the fifth digit (v) articulates on a more proximal plane than the others, i. e. nearer the trunk, as in most other Plesiosauri. In the same skeleton the pectoral limb equals seventeen of the middle cervical vertebrae in length ; J in PL macrocejjhalus it equals sixteen of these vertebras, in PL rufjosus it equals fifteen. * ' Geological Transactions,' 2n(l series, vol. v, pi. 43. t lb. X The artist has drawn the outline of the limb-bones, figs. 2 and .'5, on a larger scale than that of the skeleton, in Tab. I. 12 BRITISH FOSSIL REPTILES. Plesiosaurus homalospondylus (Tabs. V — VIII). In the year 1842 I examined, in the Museum at Whitby, Yorkshire, a collection of Piesiosaurian vertebrae, which had been taken out of a heap of rubbish from the old alum works carried on in the upper Alum Shale — a part of the Liassic series on that coast, characterised by the Ammonites lieteropliyllus. Sow. The vertebrae were divisible into two groups, indicative of two species of Vlesiomurus. Of one kind there was a series of sixteen consecutive cervical vertebrfe, charac- terised by the unusual concavity of the terminal articular surfaces of the centrum. On making a section of two of these vertebrae cemented by the matrix in their natural state of co-adaptation, the margins of the opposed articular surfaces were two lines apart, showing the thickness of the inter-articular connecting ligamentous substance at that part, while the middle of the articular surfaces left an interval of eleven lines, thus approaching the ichthyosaurian type of vertebral union. The following were dimensions of the centrum of these cervicals. In. lines. Length 19 Breadth of articular surfaces . . . . . . Ill Height of ditto 1 10 The inferior surface of the centrum showed a median longitudinal convex ridge between the two w'ide elliptical venous foramina. I named the species indicated by these vertebrae Plesiosaurus ccelospondylus* in reference to the hollow terminal articular surfaces. I hope to have, at a future opportunity, further means of illustrating this species. The second series of vertebrae presented almost flat articular surfaces of the centrum (Tab. V, figs. 3 and C) ; the inferior surface w^as devoid of a median ridge, or had only a slight rising (fig. 4, c) between the venous foramina, which were smaller and more narrowly elliptical (ib., figs. 4 and 7) than in PI. ccelospondylus ; the middle of the surface was bounded laterally by the costal surfaces (ib., pi), and was nearly flattened, being very slightly concave, both lengthwise and transversely. The costal surface is of a narrow elliptical form, with the long axis parallel with that of the centrum; the dividing line or fissure is not conspicuous ; it is situated, as usual, rather nearer the back than the front end of the centrum (Tab. V, fig. 2, ^0 ; and a space more than twice its vertical diameter intervenes between it and the neurapophysis (ib., np), or * KotXos, hollow, oXovlvXot, vertebra. LIASSIC PLESIOSAURS. 13 neurapophysial surface. A low, longitudinal rising, or obtuse ridge, traverses the free surface of the side of the centrum midway between the pleur- and neur-apophysial articular surfaces. The following were dimensions of the centrum of a cervical vertebrae answering, or nearly so, in position, to that of the Plesiosaurus cmlospondylus selected for measure- ment— In. lines. Length .......... 2 0 Breadth of the articular surface 1 10 Height of ditto 16 These dimensions showed the greater proportional length of the cervicals of the present species ; and, concurring with the more obvious difference in this shape of the terminal articular surfaces, I thereupon devised the name of Plesiosaurus homalo- spondi/his* indicative of the even or level character of those surfaces, for the species so characterised. I have subsequently received several additional vertebral evidences of both these species of the Upper Lias, or Alum Shale of Whitby, and, finally, have had the opportunity of studying two almost entire skeletons of the Plesiosaurus homaJosponddus from that locality, one of which (Tab. VIII) is now in the Museum of the Philosophical Society of York, and the other (Tab. V) has been purchased by the Trustees of the British Museum, where it is now exhibited in the Geological Department. Both of these specimens exhibit the striking character of the genus Plesiosaurus in a maximised degree, viz., in the length of the neck and the smallness of the head. I propose, first, to describe the specimen in the British Museum, Tab. V. This specimen gives indications of the same conditions of interment in its matrix, and of the operation of subsequent gradual pressure, as that of the species last described, from the lias of another part of the kingdom. It has sunk into the mud, which afterwards became petrified, either prone or supine ; for I have been unable to obtain evidence as to whether the present exposed part of the skeleton was wrought out from the upper or under surface of the block, as re- moved from the quarry ; but we may assume the former, and consider that the animal was originally imbedded with the upper or dorsal surface toward the observer. Both fore and hind paddles were outstretched at right angles with the axis of the trunk, but only their proximal bones or segments have been preserved. The skull and cervical vertebrae have maintained their original position. At the base of the neck, where the neural spines, from their height and breadth, began to afford a surface upon which the dislocating force could operate, they have begun to yield toward the left * 'O/ioXo's, planus ; aiziivhvKos, vertebra. 14 BRITISH FOSSIL REPTILES. side, and, in the dorsal region, d i — co, are turned flat in that direction. At the base of the tail, where these flattened surfaces again become diminished in extent, the vertebrae gradually resume tlieir vertical or prone position, the summits of the spines being uppermost, as far as the seventieth (counting from the head), be)fond which some dozen of the terminal caudals are jumbled together in an irregular group, as if that part of the carcass, supporting perhaps a caudal expanse of integument or fin, had been subject to some disturbing influence prior to complete imbedding in the matrix. I conclude that this partial rotation of the dorsal series took place before the petrifaction of the bones and bed ; because the ribs of the right side have slipped from their attachments to the diapophyses, in a degree corresponding with the extent of the rotation. For, had they been cemented in their natui*al connections by the Lias stone, i. c, after the petrifaction of the mud, and prior to the operation of the extraneous pressure, they might have been expected to have been bent or broken, when pressed into the same plane with the neural spines, without any slipping from their previous joints ; whereas this dislocation implies a rotting away of the articular ligaments, and a certain yielding of the surrounding bed. The chief characteristics of the skeleton of the Plesiosaurus homalosjjondylus are, the length of the neck, the height and breadth of the dorsal and contiguous cervical and caudal spines, with the smallness of the head. The length of the neck is due both to the number of vertebrae — thirty-eight, and to their proportionate length individually, and chiefly to the latter character, as compared with Plesiosaurus (Jolichodeirus (Tab. I). I caused to be carefully removed from the matrix of the present skeleton the thirteenth and fourteenth (Tab. V, figs. 2 — 4) of these instructive vertebrae, the length of the centrum in which agreed with that on which I had made notes and drawings in 1842. They corresponded in every other particular with these vertebrae. The low, longitudinal ridge or rising (Tab. V, figs. 2, 5, r) on the side of the centrum may be traced throughout the neck. Fig. 7, Tab. V, gives a view of the under surface of the eighth cervical vertebra ; fig. G gives an end view, and fig. 5 a side view of the centrum of the third cervical vertebra, all of the natural size. The specific characters are well exemplified in these, which may be profitably compared with the figures of the corresponding vertebrae of the Plesiosaurus planus, in a former Volume,* as exemplifying the degree in which vertebral characters are developed in the diS"erent species of the genus. The cervical ribs, as indicated by the articular surface (Tab. V, figs. 2, 7, pi), are of small size in proportion to the rest of the vertebra, until about the thirtieth, in which the transverse outstanding part of the stem is two inches three lines in length, and the longitudinal part two inches six lines. In the thirty-fourth vertebra this has attained * Volume I, ' British Fossil Reptiles ;' Supplement No. II to the ' Fossil Reptilia of the Cretaceous Formations.' LIASSIC PLESIOSAURS. 15 a length of four inches; the production, anterior to the transverse stem, being nearly an inch in length. In the thirty -fifth vertebra (Tab. VII, c 35) the costal surface pro- jects, the rib begins to ascend, the anterior production to shorten, the posterior one to lengthen. In the thirty-seventh (ib., 37) the rib is supported in equal proportions by the centrum and by a diapophysial growth of the neurapophysis. In the thirty - eighth (c, 3S) the rib has passed almost wholly upon the diapophysis, and has assumed a simple rib-like character, slightly bent, with a length of six inches. In the thirty- ninth vertebra (ib., D 1) the transit from centrum to neurapophysis, np, is complete, denoting the first of the dorsal series. In the second dorsal vertebra of the present skeleton the rib has slipped forward from its joint, d. In the forty-third (Tab. V, fig. 1) it is depressed an inch below the diapophysis. In the forty-sixth to the fiftieth vertebree the heads of the ribs lie beneath the centrums, and the side view of the whole of those vertebrae is obtained. In the succeeding dorsals the ribs gradually approxi- mate their suspending processes, and have resumed their articulation at the twentieth dorsal, or the fifty-eighth vertebra, counting from the atlas. The ribs of the forty-seventh to the fifty-first vertebrae are from sixteen to seventeen inches in length ; they are the longest of the series. The articular head presents a diameter of one inch and a half; the anterior surface is convex transversely; the outer part of the posterior surface is rather concave in the same direction, so that the outer margin of the proximal half of the rib, to near its head, presents the cha- racter of an obtuse rim or ridge. They gradually decrease in size as the vertebrae recede in position from the fiftieth ; and, at the sixtieth, are reduced to a length of four inches : this and the two succeeding ribs seem to have become anchylosed to the diapophyses. In the sixty-second vertebrae the rib suddenly augments in thickness, extends its articulation downward upon the centrum, and represents a sacral vertebra (Tab. I, s). That of the sixty-first vertebra is somewhat less thick, but it may have assisted in affording attachments to the ilium (ib., 02), the proximal end of which bone is in contiguity with the converging terminations of the ribs of the sixty-first and sixty-second vertebrae. The anchylosed condition, with shortening of the caudal ribs, or pleurapophyses, give them the usual chai'acter of transverse processes in the caudal region. The neural spines, thin and antero-posteriorly extended in the neck (Tab. V, fig. 2, a, 2, ns), have* been more or less broken away, in the operation of exposing the specimen, from the anterior three fourths of the vertebra? in that region. Their height gradually decreased as they approached the head and receded from that of the thirtieth vertebra (ib., 30), which rises four inches from the summit to the neural arch, having a fore-and-aft diameter of two inches three lines, and a thickness of three lines. The former diameter is least a little above the origin of the spine, and gradually increases toward the summit, where the spines are in contact. In the thirty-third vertebra the neural spine is five inches in length, and its breadth of two inches three 16 BRITISH FOSSIL REPTILES. lines is maintained through nearly the whole of that length, in corresponding* close contact with the contiguous spines. In the thirty-seventh vertebra (Tab. VI [, 37) the length of the neural spine is five and a half inches ; it has a little increased in thick- ness ; the fore-and-aft diameter continues the same. In the second dorsal the neural spine is six inches four lines in length, with a thickness of six lines. These dimensions are continued to the fifty-eighth vertebra, save that, in the posterior half of the dorsal series, the spines have less fore-and-aft breadth at their proximal third, and leave cor- respondingly wider intervals ; they are in contact at their more expanded distal portions. From the fifty-eighth vertebra they gradually decrease in length to the sixty-second, or sacral vertebra, showing a height of less than four inches, with a ter- minal fore-and-aft extent of two inches, and a thickness of six lines. They decrease in all dimensions as the caudals recede from the trunk, and most so in fore-and-aft extent, leaving wider interspaces ; by which character, with the higher position on the centrum, and anchylosed condition of the pleurapophyses, a caudal vertebra may be distinguished from a cervical of similar size. The caudal centrums are also thicker in proportion to their length, and the under surface, if exposed, would doubtless also yield the character of the hfEmapophysial pits. The dorsal diapophyses progressively increase from the first (Tab. VII, d 1, 2, d), and attain, at the fifth dorsal vertebra (Tab. V, fig. 1), a length of two inches three lines along the upper border. The rib-surface is cut from above downward and inward, shortening the under extent of the process. A low ridge is continued from the pos- terior angle of the neurapophysis upon the back part of the diapophysis, which expands to the truncate articular surface. After the sixteenth dorsal the diapophyses gradu- ally shorten to the sacral vertebrae, where they have almost subsided. The zygapophyses in the neck (Tab. V, fig. 2, z, z') and greater part of the back are nearly horizontal, the anterior ones looking a little inward as well as upward, the posterior ones the reverse ; they are given off nearer the base of the neurapophysis tlian usual (compare Tab. V, fig. 2, -, y with Tab. Ill, fig. 4, Plesiosaurus dolichodeirus, and Tab. X, fig. 1, PL rostratus), towards the end of the back their aspect gra- dually changes; and, in the tail, the articular surface becomes almost vertical; that of the anterior ones, which are most developed and longest retained in the vertebral series, looking inward. The terminal articular surfaces of the centrum of the last dislocated caudal vertebrse are, as usual, more concave than in the neck. The development of the neural spines throughout the trunk and base of the neck is such as to impede inflection in the vertical direction. At the anterior half of the long and slender neck this bend would, indeed, take place in some degree ; but the greatest flexibility would be from side to side. The provision for the attachment of the vertebral muscles in the trunk is very great, indicative of corresponding power of regulating the movements and position of the body during the application of the LIASSIC PLESIOSAURS. 17 lengthened, slender neck, and small head, in the capture of fishes or other active marine prey. The whole framework of the trunk is singularly massive, and the character of this part of the skeleton, as shown in the specimen (Tab. V), is especially striking in con- trast with the slender neck and small head of the animal. Of the Skull (Y&h. VI). The skull (Tab. VI), from the occiput to the end of the snout, is 9 inches long;, it measures 4 inches 4 lines across the middle of the temporal depressions, 3 inches 6 lines across the occiput, which rises but 1 inch in height above the foramen magnum ; the intertemporal part, or parieto-frontal crest, rises into a sharp ridge ; the length of the temporal fossa is 2 inches 9 lines, the breadth is 2 inches. The diameter of the orbit is 1 inch 6 lines ; from the fore-part of the orbit to that of the snout is 4 inches. The elliptical nostril shows a long diameter of about 6 lines, it is situated about 8 lines in advance of the orbit, and about the same distance from its fellow. The inter-narial portions of the nasal and premaxillary bones rise into an obtuse ridge. The teeth are small, slender, slightly recurved at the fore-part of the jaw, where the enamelled crown of the longest does not exceed 10 lines. No sutural evidence of cranial structure is discernible ; tlie bones about and between the orbits show the effects of pressure. Estimating the length of the skull by that of the lower jaw, about two inches should be added to that taken from its exposed and visible part. This part of the skull (Tab. V) is susceptible of satisfactory comparison with the corresponding region of the skull in the Plesiosaurus dolichodeirus (Tab. Ill, fig. 1), the species which most resembles the Plesiosaurus homalospondylus in the length of the neck and the small proportional size of the head. By comparing Tab. Ill with Tab. VI, in which the skulls of the two species are figured of the natural size, from probably mature individuals of average size, and from the same aspect, the difference of proportion and form is such, and so obvious, that, were two skulls of existing lizards to be so contrasted, it is probable that some Erpetologists would be led to sever them more widely than by specific bounds. The composition of the cranium, the position and relative size of its principal cavities, and especially of the nostrils, the character of the dentition, are, however, so strictly Plesiosaurian in the two fossil skulls here compared, that there is no suSicient ground for encumbering the Sauropterygian group with one or two additional generic names. The skull oi Plesiosaurus homalospondi/his is longer in proportion to its breadth, more oblong in shape, more obtusely terminated anteriorly. It is possible that the skull of the Plesiosaurus dolichodeirus compared (Tab. Ill) may have suffered more horizontal pressure, but not such as to have affected its triangular shape due to the d 18 BRITISH FOSSIL REPTILES. more rapid convergence of the sides of the upper jaw to the more pointed muzzle. The temporal fossae may appear broader than natural in this crushed skull, but with due allowance this shape was square, not oblong, as in Plesiosaurus homalospondylus. The intervening parieto-frontal crest is relatively longer, and we may infer that the biting muscles were larger and more powerful in relation to the more massive propor- tions of the dentigerous parts of the jaws in Plesiosaurus homalospondylus : the orbits are relatively less ; their antero-posterior diameter is less than one fifth of the same diameter of the skull taken from the back part of the parietal (7) in PI. homalospondylus; it is more than one fifth in PI. dolichodeiriis ; the orbits are equidistant from the two extremes of this diameter in PI. homalospondylus ; they are nearer the back part of the head in PI. dolichodeirus. In PL rostratus (Tab. IX) the temporal fossae present some- what intermediate proportions between those in the two foregoing species ; but the rostral production of the maxillary part of the skull sufficiently distinguishes the cranium of P/. rostratus from that of previously known species in a comparison of- detached skulls ; whilst its greater relative size to the body more especially distin- guishes it from that in PI. homalosjmndylus or PI. dolichodeirus. In PL Hawkinsii* the longitudinal diameter of the temporal fossa exceeds the transverse diameter, but not in so great a degree as PL homalospondylus, and the upper jaw is relatively narrower than in that species. This is also the case in PL macro- ce]3halus,\ in which there is a more marked constriction of that part, anterior to the orbits, showing a tendency to the " rostral " character, which is exaggerated in PL rostratus. Pectoral and pelvic arches and limbs (Tabs. V and VIII). Of the limbs only the humeri and femora have been preserved in the skeleton (Tab. V) ; these bones show the usual form, with their respective characteristic modi- fications, as exemplified in the different contour of the anterior border, which is straight or partly convex in the humerus, and is concave in the femur. The length of the humerus is 12 inches, that of the femur 13 inches; the distal breadth is nearly the same in both, namely, 6 inches. In the right femur, the coarse fibrous texture which pervades the whole thickness of the bone is exposed. A portion of the exten- sive scapulo-coracoid arch comes into view from beneath the anterior dorsals on the right side (Tab. V, 52). The ilium (ib., 62) presents the usual form; straight, slender at its proximate end, with a slightly twisted, subcylindrical shaft, expanding to a breadth of nearly three inches at its acetabular end. The entire length of the * 'Geol. Trans.,' 2nd series, vol. v, pi. 45. t Ibid. LIASSIC PLESIOSAURS. 19 skeleton (Tab. V) is fourteen feet, which would be increased by several inches were the tail entire and outstretched. The specimen of PL homahsjmiilijlus in the Museum of the Yorkshire Philosophical Society is larger than the one in the British Museum, but has been lithographed on a smaller scale in Tab. VIII; it measures IG feet 6 inches in total length. It lies in a somewhat similar posture to that in the British Museum, but with the long and slender neck and anterior dorsals bent so as to give a concavity to the dorsal con- tour of the animal ; the caudal vertebrae, which are better preserved, are also bent in the same direction, and all the vertebras follow in their consecutive undisturbed jiixta-position in both skeletons. The numbers of the vertebrae in the cervical and dorsal series respectively appear to be the same. The diapophysis has got entire possession of the rib at the fortieth vertebra, counting from the head ; and the costal surface begins, with its process, to sink again upon the centrum, at the sixty-seventh vertebra, which the thickness of the diapophysis indicates to be a sacral vertebra. Beyond this may be counted twenty-seven caudal vertebrae, and it is not probable that their number exceeded thirty. The cervical vertebrae show the same distinctive characters of the species which have been already defined ; the neural spine is preserved in a much greater proportion of the cervical series ; in the fifteenth cervical it shows a height of two inches, and a nearly equal antero-posterior breadth ; with a broadly truncate summit, having the angles rounded off. The vertebrae keep their proportion of length from this point to the end of the dorsal series ; they then grow shorter to the end of the tail, throughout the greater part of which the centrum is deeper, and the neural spines longer and narrower, than in the neck, indicative of the greater mass of muscle operating on the tail, and also its greater flexibility in a given extent. The costal series has suffered much more displacement and loss in the York specimen than that in the British Museum; the larger ribs are a good deal jumbled and broken in the region of the trunk or thoracic abdominal cavity, but they show the same massive character. The ischio-pubic part of the pelvis has been drawn away, at an acute angle, from the ilium and sacrum; its inner or upper surface is exposed at 63, Qh Tab. VIII. The right pelvic limb has been moved forward, with the head of the femur lying upon the lower end of the right coracoid. The right pectoral limb extends forward from near its normal place of articulation with the coracoid ; but it has been turned bodily over, showing its inner or palmar surface. The limbs of the left side are huddled in a dislo- cated and incomplete state beneath the hinder part of the trunk. The presence of both these limbs, in an excellent state of preservation, supplies the chief deficiency in the specimen in the British Museum previously described. The pectoral limb, as in PI. clolichoddms, is rather shorter than the pelvic one ; its entire length is 3 feet 8 inches, equalling sixteen vertebrae towards the base of the neck. The humerus, 13 inches in length and 7^ inches in distal breadth, is broader 20 BRITISH FOSSIL REPTILES. there, in proportion to its length, than in the PI. dolichodeirus or than in the PL rostratiis ; its anterior margin, as in the skeleton Tab. V, is more straight than in those species. The antibrachial bones (54, 55) show intermediate proportions of length and breadth between those in PI. dolichodeirus (Tab. I, fig. 2) and PI. roHtratus (Tab. IX). They present the usual characteristics of radius (54) and ulna (55) in the present genus, and they are of equal length. The hand measures two feet in length, and is somewhat longer in proportion to the arm and forearm than in the two above-cited species ; it also shows rather more breadth. The carpus consists of six bones, three in each row, and with less inequality of size. The distal bones occupy an equal breadth with those of the proximal row, and do not allow the base of the fifth metacarpal to extend backward to the proximal row, as in the species of which the carpus is figured in Tab. XIV. The bases of the five metacarpals (in Tab. VIII, 57) are on the same transverse line ; and if this specimen should truly exhibit the relative position of the bones of the pectoral fin, characteristic of the species, it adds a well-marked distinction of the PI. homalospondylus. The first, or radial, or innermost metacarpal (57), supports a short digit of three phalanges ; the second a digit of seven phalanges ; the third the same ; the fourth has a digit of six phalanges ; the fifth is obviously imperfect. The pelvic limb (Tab. VIII, G5, 69) is 3 feet 9 inches in length ; the femur (65) is 14 inches long and 7^ inches across the distal end. The tibia and fibula are respec- tively longer than their homotypes the radius and ulna ; the foot is 2 feet in length and 1\ inches in basal breadth. The tarsal bones are similar in number and arrange- ment to those of the carpus ; and as the bases of the five metatarsals (69) are in this limb also on the same transverse line, I have the greater confidence in the natural structure being here shown in both limbs, and that they thus exhibit a distinctive character, of specific value, from the other Plesiosauri described in the present Monograph. Plesiosaurus rostratus, Owen. Tabs. IX — XIII. The specimen on which this species is founded was obtained, in 1863, by Edward C. Hartsinck Day, Esq., F.G.S., from the Lower Lias at Charmouth, Dorsetshire, by whom it was transmitted to London for inspection, and it has been purchased by the Trustees of the British Museum, where it is now exposed in the gallery of Geology. It is figured, one ninth of the natural size, in Tab. IX. This skeleton, like most of the plesiosaurian ones that have come under my obser- vation, indicates the ordinary and tranquil character of the death and burial of the individual ; it has sunk entire, relaxed, and prone, with outstretched limbs, in its matrix, when this was soft and yielding ; and, as decomposition loosened the liga- LIASSIC PLESIOSAURS. 21 mentous attachments of the vertebrae and of their elements, they have yielded to external pressure or movement of the matrix, and have rotated on their axis — some of the long-spined vertebrae to the right, some to the left — with a slight displacement of the longer ribs from their attachments. The third cervical vertebra is displaced about three inches below the axis and atlas, which remain in connection with the occipital tubercles, the third to the fifteenth cervicals are prone with the spines uppermost, and the pleurapophyses in natural connection with the sides of the centrum, the lower part of which is buried in the matrix. Except a slight dislocation between the seventh and eighth, these cervicals have retained their natural sequence and relative position. As the spines grew longer and larger they offered a surface upon which the superincumbent pressure could operate, so as to rotate the vertebrae sideways ; and from the sixteenth to the twenty-eighth inclusive, they are turned half round, with the spines downward or to the left ; but all these vertebrae retain their natural mutual connections. The twenty-ninth vertebra is dislocated, exposing the anterior articular surface of the centrum ; the thirtieth has suffered fracture of its spine ; the thirty-first and thirty- second are partly bent to the left ; the thirty-third and thirty-fourth are turned with _the spines to the right side ; that of the thirty-fifth is broken from its neural arch ; the thirty-sixth to the forty-eighth vertebrae have the neural spines turned to the right, retaining almost their natural relative positions. The forty-ninth vertebra has kept the original prone position, as when imbedded ; the next ten show the side view, with the neural spines to the right; the sixty-first to the sixty-fifth are prone, but with a slight deviation of the neural spines, some to the right, some to the left ; the next six vertebrae have yielded in the opposite direction ; there is then a deeper space, equal to the extent of five vertebrae, in which there are the centrums of three vertebrae and some hsemapophyses irregularly scattered. Beyond this part the terminal caudal vertebrae resume their position and natural connections, and are preserved, seven in number, to the last. The antecedent exceptional violence shown in the caudal series has probably been due to the tugging and gnawing of some predatory animal, whilst this part of the dead and partly decomposed Plesiosaur continued to be exposed at the sea-bottom. The scapulas (51) and articular ends of the coracoids (52) appear parallel with the twenty-fifth to the twenty-seventh vertebra3, the left being rather further back than the right. Both humeri (53) have been dislocated at the shoulder-joint by super- incumbent pressure, and the articular ends of the scapulae overlap their heads. The rest of the bones of the pectoral fins have retained their natural relative position, protected by the tough, closely-fitting dermal sheath, until this slowly dissolved away. The iliac bones (fi2) lie by the sides of the forty-seventh to the fiftieth vertebrae, almost in the axis of the spine, with their proximal ends turned backward, and their acetabular end forward, having become detached from the thick, converging pleura- 22 BRITISH FOSSIL REPTILES. pophyses of llie forty-seventh and forty-eighth vertebrte (s, s) which overhe the ischium (63) on the left side of the body. The articular ends of the ischium (63) and of the pubis (64) are exposed, retaining their connection with the ilium (62) opposite the fort)'-third to the forty-seventh vertebrte on both sides. The femora (65) have been slightly dislocated forward, and part of the acetabula is thus exposed. The bones of the hind fins have preserved their natural relative positions; those of the left side, with their part of the pelvic arch, being a little more backward in position than those of the right, agreeing, in this respect, with the pectoral limbs, and indicating some general movement of the matrix as the cause of such displacement. Including the atlas and axis there are twenty-four vertebrse before that in which the pleurapophyses have risen, to articulate wholly with the diapophyses (Tab. XII, d). At the forty-fifth vertebra the rib again begins to articulate with the centrum ; in the forty-sixth the parapophysis forms the lower half of the costal surface; in the forty- seventh it forms a larger proportion, and the whole costal surface is here suddenly increased in size, giving attacliment to a short, slightly bent pleurapophysis of correspondingly and abruptly increased thickness ; that of the forty-eighth vertebra is thicker and straighter, and, as the preceding riblet inclines towards its extremity, 1 conclude that their thick, abrupt, digital ends were ligamentously connected with the iliac bone, and that they therefore may be regarded as sacral vertebrae (Tab. IX, s, s). The remaining vertebrae, from the forty-ninth to the eighty-fourth, will be caudal ; thus there may be reckoned 24 cervical, 24 dorsal, 2 sacral, and 34 caudal vertebrae, in the present species. Perhaps the two vertebras antecedent to the sacral, in which the centrum shows part of the costal surface, might be regarded as lumbar vertebrae. The total length of the vertebral column, from the third cervical to the last caudal, following its slight undulations, is 9 feet 9 inches. The skull, from the hind end of the mandible to the fore end of the symphysis, or snout, is 1 foot 11 inches. The first five or six cervicals, from the third, are more or less obscured by pyritic matter ; their neural spines show intervals of from three to six hues ; the upper margin of the spine rises obliquely from before backward, with the angle rounded off; it is thickest at the middle part, where it measures two lines ; that of the fourth vertebra has a fore-and-aft diameter of seven lines, the same diameter of the ninth is one inch. The pleurapophyses of the tenth vertebra are about an inch in length, with a subcylindrical body, bent obliquely backward, and slightly tapering to an obtuse end. In the eleventh vertebra, the centrum of which is an inch in length, about five lines of free surface intervene between the costal and neurapophysial articulations. From the pleurapophysis to the summit of the neural spine it measures 2 inches 5 lines. At the twelfth cervical the pleurapophyses begin to send forward the process which marks what may be termed the neck or pedicle of the cervical rib. LIASSIC PLESIOSAURS. 23 At the fourteenth cervical the length of the centrum is 1 inch 5 lines ; that of its pleurapophysis is 1 inch 9 lines ; the fore-and-aft extent of the base of the neural spine is 1 inch 2 lines ; the height of the spine is 1 inch 6 lines, and its thickness is 3 lines. The total height of the vertebra is 4 inches. These dimensions are gained by gradual increase from the tenth vertebra. In the nineteenth cervical the length of the centrum is 1 inch 6 lines, the space between the pleur- (ib. p/)) and neur- {ih.np) apophysial surfaces is 7 lines. From the lower part of the centrum to the summit of the neural spine ( ns) is 5 inches ; the length of the pleurapophysis is 2 inches. The P/ef. Frc. 6. Neural arch, front view. Caudal vertebra, side view. Nat. size. Nat. size. In the pinnigerous part of the column the neural arches (Cut, fig. C) gradually lose breadth, gain length, and still more gradually lose height, the spine always predominating in the fore-and-aft diameter ; the summit equals the centrum in that extent, and the neural spines of contiguous vertebrae touch, but do not overlap. The vertical contour of the midpart of the centrum, above and below a more promi- nent convex part, is slightly concave. The borders show the moderate convexity or thickening, relating to the compressed characters of the pinnigerous caudals. The neurapophyses in the basicaudal region seem not to have coalesced above, find the broad, laterally-impressed, and backwardly-produced part, simulates the half of a truncated neural spine. The short, straight, and inferiorly-situated pleurapophyses con- tinue to be developed to near the tail-bend. The shorter hsemapophyses (Cut, fig. 6) are continued from the pinnigerous caudals to within a third part of the pointed tail- bend. In the skull of Ic/iiJ/j/osarus intermedins the following characters may be noted. The under surface of the basioccipital is but slightly excavated anterior to the condyle, and the ' foramen parietale ' is almost wholly in a raised part of the hind end of the inter- 72 BRITISH FOSSIL REPTILES. frontal suture. The maxillary is separated from the external nostril by a junction of the premaxillary with the lacrymal. The sclerotic plates are often from fifteen to seventeen in number. The surangular is deeper, and forms a larger proportion of the outer surface of the hind half of the mandibular ramus than the angular. It terminates or disappears at the usual point between the deutary and spleuial, in advance of the angular ; it develops on its upper border a small but well-marked corouoid angle bounding anteriorly the con- cavity under the hind part of the orbit. Beneath this angle or process begins the neuro- vascular groove, which extends, gradually shallowing, a short way forward. The splenial element begins to show at the lower margin of the ramus about the mid-length of the angular; it unites with its fellow to form the hinder two thirds of tbe symphysis mandi- bulae (Tab. XXVI, fig. 2, 32'). The articular piece is brought into view by its partial dislocation backward in the right ramus of the subject of fig. 1, Tab. XXVI. The stem of the episternum equals in length one half of the cross-bar; in receding therefrom it slightly expands and becomes flattened. In one young specimen a median cleft extended a short way forward from the end. The clavicles are distinct, long, and less strong than in Ich. communis, feebly bent, with the concavity behind and within, gradually narrowing to each end, and having a sutural surface beneath and behind, some way along each end ; the shorter one engrains with the episternal cross-bar, the longer one with the fore part of the scapula. Of this bone the fore border is straight, the hind one concave through the backward production of the humeral joint ; near this the outer surface is shghtly excavated. In the coracoid the humero-scapular articulation is of less relative extent than in IcMliyosaurus communis. The anterior notch is broader, but is as deep as the hinder one ; the intervening tract, or neck, is relatively larger than in Icli. platyodon or Iclb. tenuirostris. The present species has afforded some ground of the restoration outlined in fig. 1, Tab. XXIX. The humerus is marked 53 ; the radius 54 ; the ulna 55. The fore border of the radius is entire, not notched. The antibrachium supports, as usual, three ossicles (56), followed by four rather smaller (56')- Regarding these as carpals, they support four metacarpal ossicles (57—57'), from which are continued five longitudinal series of progressively diminishing phalanges. The anterior or radial metacarpal (57') supports two digits, of which the radial may be symbolised as I, the next series as digit 11. The three other series. III, IV, V, are supported by their respective meta- carpals. Each digit consists of numerous phalanges, progressively decreasing to the end. A series of small supplementary ossicles is applied to the radial border of digit I, and to the ulnar border of the digit V, so that seven ossicles may be counted in the same transverse line along the middle third or fourth of the series. The number of phalanges is, however, less than in Ich. communis, and the fin is relatively narrower. The characters of the fore paddle above defined are well shown in specimens from the Lias of Lyme Regis and of Street, in the British Museum. LIASSIC ICHTHYOSAURS. 73 In the pelvic fin the femur is longer in proportion to its breadth, and the distal expansion is relatively greater than in the humeinis. The tibial ossicle of the three tarsals has an emarginate tibial border ; the corresponding ossicle or phalanx of the second and third series shows the same character. d. Ichthyosaurus platyodon, Ch. Tab. XX, figs. 4, 4' ; Tab. XXVII, figs. 1, 2, 3. The skull of Ichthyosaurus platyodon (Tab. XXVII, fig. 2), is somewhat longer in proportion to the trunk than in Ich. communis. Taking as the trunk the extent of the vertebral column to the pelvic arch, such extent includes, in the subject of fig. 1, one. length and a half of the skull, while in Ich. communis (Tab. XXIV, fig. 1) it includes rather less than two lengths, and, in Ich. intermedins (Tab. XXVI, fig. 1) rather more. The skull ol Ich. platyodon is longer in proportion to its breadth than in Ich. inter- medius. The jaws are stronger from the greater relative depth of the mandible and the less gradual attenuation to the rostral extremity. The orbit is of a full elliptic shape, with less approach to the circular, than in Ich. breviceps (Tab. XXV, fig. 1). It is relatively less than in the long- and slender-snouted kinds. The length of the rostrum anterior to the orbit is three and a half times the longitudinal diameter of that cavity. The osseous circumpupillary ring includes thirteen sclerotic plates in the subject of Tab. XXVII, fig. 2. The surangular (ib., fig. 2, 30) disappears betw^een the dentary (33) and angular (31). This element similarly disappears in a pointed form between the dentary (33) and splenial (32), beyond the midlength of the ramus. A dental characteristic of the present species is that which suggested to Conybeare the name platyodon ; the smooth enamelled crown (Tab. XX, figs, 4, 4') being subcom- pressed, sharp-edged, and pointed ; the longitudinal grooves of the tumid cement-clad root are soon lost upon the coronal base. I have counted 45 — 45 of these teeth in an upper jaw, and 40 — 40 in a lower jaw, and have noted that the crowns are more often snapped ofi" than in the smaller species, which may be indicative of the greater violence with which they have been used. From the occiput to the iliac bones there are forty-five vertebr£e ; thence to near the end of the tail may be counted seventy-five vertebrae ; the total number in the skeleton probably somewhat exceeded 120. One of these vertebrae, from the hinder half of the abdomen, is figured (in the inverted position with the neural surface downwards) in Home's Memoir of 1816.^ The neui'al spines are thicker, shorter, and more rounded superiorly than in Ich. intermedius or Ich. commmiis. The zygapophyses, especially the ^ ' Phil. Trans.,' mdcccxvi, pi. siv. 10 74 BRITISH FOSSIL REPTILES. anterior ones, are well developed, and the vertebrae of the trunk and basal moiety of the tail are strongly interlocked, though admitting some inflection. The ribs increase in length to the twenty-fifth pair ; at the thirtieth pair they begin to shorten gradually, and, after the fortieth pair, more suddenly, becoming nearly straight at the forty-fourth pair. Thence they are continued like long transverse processes articulated by a simple head to the single di-parapophysis, as far as the hundredth vertebra. The scapula (Tab. XXVII, fig. 3, 5i) has a relatively broader humeral end than usual. It is preserved with the corresponding clavicle (ib., as) in a portion of a huge skeleton of the present species in the British Museum. The coracoid has a relatively larger or longer scapulo-humeral surface than in Ich. intermedins, and has a narrower and deeper anterior notch or emargination than in Ich. communis; the ento-sternal margin is rather thicker than usual. I have noted a specimen of this bone from Lyme Regis, of which the long diameter was 8 inches 4 lines, the short diameter 6 inches.^ The humerus is notable for its breadth, especially distally, compared with its length- The proximal rounded end, or ' head,' is tuberculate at its circumference, indicative of powerful ligamentous attachments to the scapulo-coracoid joint. The fore margin is more concave than usual. This latter character is still more marked in the radius, which, with the ulna, presents the generic shortness and flatness, with a slight excess of breadth, as compared with most other species. The anterior emargination is present also in the radio-carpal bone (Tab. XXVII, fig. 1, 54), and in the corresponding one in the following series. The next ossicle presents the common pentagonal form. Not more than three series of digital bones are preserved in the subject of figure 1, Tab. XXVII. A few supplemental ossicles are preserved at the radial border beyond the middle of the fin- framework. I have not found evidence of a greater number of pectoral digits in any remains of the present species. It seems to have been characterised by long and narrow, but powerful fore paddles. In the pelvic bones the ilium (Tab. XXVII, fig. 1, C2) presents a straight, flattened, slender form. The ischium (ib., G3) is remarkable for its breadth, especially at its medial end. The pubis (ib., 04) is less expanded there ; its anterior border is straight. The femur (ib., 60) is longer in proportion to its breadth than the humerus ; its proximal end shows a large depression, probably for the insertion of a stout ligament. The tibia (ib., 66) presents an anterior emargination, as in the radius (54) ; the same character is repeated in the two succeeding ossicles at the same margin of the fin- framework. Here, also, but three digital series are preserved, with a few small supplemental ossicles along the fibular border of the fin. The disposition of the distal ossicles in both pairs indicates that the ligamentous or fibro-cartilaginous uniting medium of their framework may have been more abundant than usual, allowing greater flexibility of the terminal part of the long and narrow 1 ' Report,' ut supra, 1839, p. 114. LIASSIC ICHTHYOSAURS. 75 paddles of IchlJii/osaurus jilafi/odon. The least incomplete skeleton of this huge species in the British Museum, the subject of Tab. XXVII, fig. 1, is from an individual of about 20 feet in length ; but portions of others — the skull, for example, which may be seen at the Geological Society's Rooms at Burlington House — indicate a total length of the individual so represented, of at least 30 feet. The Lias of the Valley of Lyme Regis is the chief depository of Ich. platyodon, but its remains are pretty widely distributed in the same Mesozoic zone. They have been found in the Lias of Glastonbury, of Bristol, of Scarborough, of Whitby, and of Bitton in Gloucestershire. The Ammonites associated with the bones of the subject of Tab. XXVII, fig. 1, are of the species Arietites semicostatiis, characteristic of the greyish limestone (Lower Lias) of Lyme Regis. e. Ichthyosaurus lonchiodon/ Oto. Tab. XX, figs. 6, 0'; Tab. XXVII, figs. 4 — 7. This species, which appears to have attained a bulk second only to that of the Ich. platyodon, differs in the shape and smaller relative size of the teeth (Tab. XX, figs. 6, 6'). They are more slender in proportion to their length than in Ich. communis (ib., figs. 5, 5'), and are straighter than in Ich. tenuirostris. Their base is cylindrical, less ventricose than in Ich. plati/odo7i (ib., fig. 4'), and more finely and regularly fluted than in Ich. communis. A smooth boundary divides the base from the enamelled crown, and this is traversed by fine longitudinal grooves converging to the apex. The transverse section of its base is nearly circular ; it tapers gradually to the apex, which is nearer the posterior line or contour than the axis of the tooth. The vertebral centrum has a greater proportional fore-and-aft extent than in Ich. platyodon ; the neural arch and spine have a less vertical, in proportion to the fore-and-aft, extent (Tab. XXVII, fig. 6) than in Ich. communis or in Ich. tenuirostris. Forty-five of these vertebrae may be counted between the occiput and the pelvis; and as many beyond may be made out in the subject of fig. 4, Tab. XXVII, as brings the number up to 120, but the tail is incomplete. The length of the rostrum anterior to the orbit (Tab. XXVII, fig. 5) includes little more than three longitudinal diameters of that cavity ; it is thus relatively shorter than in Ich. platyodon, and it is relatively more slender than in Icli. intermedius. The mandible, however, does not partake of this proportion, but is nearly as deep and strong, rela- tively, as in Ich. platyodon (ib., fig. 2). The surangular (30) extends farther forward than the angular (31) ; both disappear in the usual pointed form. The nostril is divided from the maxillary by union of the premaxillary with the lacrymal. The pair of hyoid elements (cerato-hyals) are preserved in situ in the subject of fig. 5 ; they are cylindrical, 1 'Report,' ut supra, p. 116. 76 BRITISH FOSSIL REPTILES. almost straight, truncate at each end, which very sHghtly exceeds in thickness the rest of the bone. Each is about one fifth the length of the mandibular ramus. The coracoid (Tab. XXVII, fig. 7) shows a deep anterior notch, with only a feeble concave outline at the corresponding part of the hind border of this bone. The clavicle is applied and suturally attached to the lower half of the fore border of the scapula. That border is nearly straight; the hinder one is concave through the backward pro- duction of the thickened lower end to contribute to the articular surface for the humerus. The anterior border of the radius and succeeding ossicle is emarginate. The chief phalanges of but three digits are preserved in the subject of Tab. XXVII, fig. 4. The character of the pectoral fin of the present species is probably rightly indicated in the main; and in such essentially tridactyle character Ichthyosaurus lonchiodon may a^ree with Ich. platyodon. The lower or distal ends of ischium and pubis seem to be equally expanded ; both bones are broader than the ilium. The ventral fin has been dislocated and bent back- ward close to the spine. The homotypal ossicles show the same emargination as in the pectoral fin. The neural spines are relatively short (ib., fig. 6). This species has hitherto been found only at Lyme Regis ; it appears to be a rarer Liassic Ichthyosaur than the three preceding ones. The skeleton above described was discovered by Miss Mary Anning, to whom the discovery and extrication of many rare and interesting fossils of the Lias of this locality are due. /. Ichthyosaurus longifrons, Oio. Tab. XIX, figs. 1 — 5 ; Tab. XX, fig. 1 ; Tab. XXI, fig. 1 ; Tab. XXII, fig. 1 ; Tab. XXIII, figs. 2—5. The characters of this species are those in which the subjects of the above plates and of the general description of the cranial organisation of the genus differ in the specific modifications referred to in other sections of the present Chapter. From the Upper Lias, with remains of Ammonites bifrons, of the Cotteswold Hills. The more complete cranial specimen from the same Liassic zone at Curcy, Normandy, has afforded the subjects of the figures in the plates above cited. g. IcHTHyosAUiius LATiFRONS, Koii. Tab. XIX, fig. C ; Tab. XXIII, fig. 1. In the year 1825 Mr. Konig, Keeper of the Department of Mineralogy, British LIASSIC ICHTHYOSAURS. 77 Museum, published a series of lithographs of fossils, in a folio form, with brief notices of the subjects, reaching to number 100. Beyond this, names alone are given at the foot of each plate, and No. 250 of plate xix bears that of Ichthyosaurus latlfrons. It is a very reduced view of a mutilated skull and portion of the vertebral column, with some vertebrae outlined of the natural size. The specimen is stated to have been obtained from Lyme Regis. Of this specimen a view of the upper surface of the skull is here given of the natural size, in Tab. XXIII, fig. 1. The upper apertures of the temporal fossae (t, t) are nearly equilaterally triangular in form, the base being external and slightly exceeding the sides, which converge to the cranium proper. The sagittal suture (7) persists, but the frontal one is obliterated, and the midfrontals (n) constitute a single symmetrical bone, which is moderately convex both lengthwise and crosswise. The " foramen parietale," of a full ovate figure, is formed wholly by the frontal, the apex alone forming the beginning of the parietal suture. The postfrontal (12) extends upon the coronal suture, and overlaps part of both parietal and frontal bones. The nasal (15) overlaps the fore part of the frontal, and divides that bone from the prefrontal (14). The sides of the skull converge rapidly to the beginning of the snout. A breadth of cranium across the temporal apertures of 7 inches is reduced to 2 inches anterior to the nostrils (ib, »). Additional characters of the present species are afforded by a second specimen from Lyme Regis acquired by the British Museum. It is a skeleton, lacking both ends, but including the trunk, with chief part of the skull and basal portion of the tail, the total length being 4 feet 10 inches. From the occipital condyle to the pelvis it measures 2 feet 6 inches ; the length of the preserved portion of the skull is 1 foot 4 inches. In this specimen is instructively shown the sudden slope by which the broad cranium descends into the rostrum. The orbit is correspondingly large, its vertical diameter is 5 inches, its antero-posterior one is 4J inches. The sclerosteal circle is composed of correspondingly large plates, of which seven are preserved. The vertical diameter of the base of the rostrum, taken across the middle of the nostril, is but 1 inch .3 lines. The premaxillary is impressed by a longitudinal groove, running about three lines above the alveolar border. The teeth are long, slender, slightly recurved ; six or seven may be counted in an extent of two inches of the alveolar groove. The length of the crown of the best preserved is 6 lines, its basal breadth being 1 line. In the composition of the mandible the angular element is unusually short ; it disappears parallel with the hind border of the orbit. On the other hand, the siir- angular is longer than usual, but its chief character is the continuation of the forwardly- directed nervo-vascular foramen, usually present below the hind border of the orbit, into a groove continued forward, towards the lower border of the ramus, and terminating at a vertical line dropped about a nostril's extent in advance of that opening. A second 78 BRITISH FOSSIL REPTILES. and narrower longitudinal groove extends along the dentary about two lines below the alveolar border. The length of the jaws cannot be determined in this or in the typical specimen, their fore end being broken off, but sufficient remains to indicate that Ich. laticeps belonged to the long and slender-jawed species, with small and slender teeth to match ; and associated, as in Ich. temiirostris, with a powerful fore paddle, supported by compa- ratively few and relatively large phalanges. The proportion of the longitudinal to the transverse diameter of the vertebral cen- trums is contrasted with those of the Ich. braclipjwndi/his in Tab. XXIX, fig. 6. The neural spines of the dorsal vertebrae are relatively short and with distinct intervals. Thirty-eight are preserved between the scapula and ilium. The scapula is characterised by the greater relative expanse of its articular end in comparison with the breadth of the body, resembling in this respect that of Ich. 2)lafyodon. The coracoid has a deep anterior emargination, and a shallow posterior one ; the antero-posterior breadth of this bone is 4 inches, the transverse extent is 3 inches. The left coracoid, with the corresponding humerus and a few paddle-bones, have been puslied dextrad and appear beneath the right coracoid, from which the corresponding paddle has been removed. The pelvic bones of the right side are well shown. The ilium, 2 inches 4 lines in length, is directed obliquely backward and downward ; its upper end is one inch in breadth. The pubis, 2 inches 3 lines in length, has a distal breadth of 11 lines ; its fore border is almost straight. The ischium, 2 inches in length, has a distal breadth of 1 inch 3 lines. The margin towards the pubis is more concave than the opposed one of the latter bone. The hind border of the ischium is moderately concave. Compared with the same bones in Ichthyosaurus communis the pelvic elements are more robust. Of the structure of the appendage of the pectoral and pelvic arches, or fins, I have not. as yet, obtained satisfactory evidence. The fossils which have served in the present section of the Chapter have been obtained from the Lias of Lyme Regis and Charmouth. h. Ichthyosaurus acutirostris,^ Oio. Tab. XXIV, fig. 2. This species is so named from the slender, sharp-pointed form of the snout, unaccom- panied by such proportions of length as characterise the Ichthyosaurus tenuirostris and Ich. lonyirostris (ib., fig. 3). 1 'Report,' ut supra, p. 116. LIASSIC ICHTHYOSAURS. 79 The length or fore-and-aft diameter of the orbit in the subject of the above plate is G inches, that of the part of the skull anterior thereto is 18 inches 7 lines. Both upper and lower jaws are impressed by a deep and narrow longitudinal groove near to and parallel with the alveolar border. The osseous sclerotic part of the eyeball occupies about two thirds of the long diameter of the orbit. The teeth are intermediate in character and in number between those of Ick. inter- mediiis and Ich. tenuirostris. The few trunk-ribs preserved in my present subject are slender, rounded, ungrooved, and have a feebly-produced anterior margin along their proximal fourth. The mesial border of the coracoid is sinuous, the articular surface for the episternum being better defined and more tumid than usual. The surface of the lateral or outer border for articulating with the scapula and humerus is strongly developed. The length of the coracoid in the specimen described is 7 inches, its breadth 7^ inches. The humerus is 7 inches in length and 5^ inches in breadth at the distal end. Of the fore paddle three digital series and a small portion of a fourth are preserved, but in such juxtaposition as to leave little doubt as to any considerable part of the fin-bones being lost. These, along the fore or radial border, including the radius, radio-carpal, and succeeding ossicle, are emarginate ; the rest have that border entire and moderately convex. After the fifth ossicle from the humerus the subquadrate merges into the trans- versely oval form. From the inclination of the radial digit toward the middle of the fin, a bifurcation is indicated at about the ninth ossicle from the humerus, and an irregular scattered series of small, full-elliptic, and circular bonelets, may be interpreted as an additional digit to the three normal ones, the more direct continuation of digit ii now extending down the centre of the bony paddle. An irregular ulnar series like that on the radial side is partially shown. If there should be lack of osseous evidence of the breadth of the fore fin there is less of its length, which seems to have been two thirds that of the skull, and this is due rather to the size than the number of phalanges. The preserved basal portion of the left paddle repeats the character of the corresponding part of the right one above described, and so far confirms the inference as to the specific character of the fins associated with the acuminate one of the skull. The fossils on which the above species of Ichthyosaurs is founded are from the Lias of Whitby, Yorkshire. i. Ichthyosaurus tenuirostris, Cb. Tab. XX, fig. 8 ; Tab. XXI, fig. 3 ; Tab. XXVIII, figs. 1— G. The characters which, in 1822,^ strikingly distinguished the present from the then 1 ' Trans, of the Geological Society,' 2nd series, vol. i. 80 BRITISH FOSSIL REPTILES. determined species of Ichthyosaurus, were the great length and slenderness of the jaw- bones, suggesting the proportions of those of the Crocodilian gharrial ; and which, in combination with the large orbits and low broad cranium, gave to the skull a resemblance to that of a gigantic woodcock {Scolopax), with a bill armed with teeth. The length of the snout is chiefly due to the prolongation of the premaxillaries (Tab. XXVIII, fig. 2, 22) and dentaries (ib., 33). The length of the skull anterior to the orbit is somewhat less than four times the antero-posterior diameter of that cavity. The parietals retain their sagittal suture, the fore part of which recedes in a greater proportion than usual to contribute to the foramen parietale. Their posterior bifurcation is applied, in the occipital region, to the super-occipital, which is broad and arched. The outer end of each parietal prong is obliquely truncate for the suture with the mastoid (ib., fig. 2). This, as usual, forms the blunt but prominent supero-lateral angles of the hind part of the cranium. The midfrontals (ib., 11) retain their suture, their lateral border articulates with the prefrontal (14) and the superorbital, or a forward extension of the postfrontal. This element articulates with the lateral border of the parietal and combines with that bone in forming the upper three fourths of the temporal fossa, the lower boundary of that cavity being completed by the mastoid. The orbit is bounded behind chiefly by the postorbital. The malar is unusually long, extends from the lacrymal (73) to form the rest of the anterior boundary of the orbit ; then, continuing to circumscribe it below, the malar curves with a rather abrupt bend upwards to join the postorbital, and by an oblique suture the zygomatic. This bone is continued obliquely backward to contribute to the external meatus auditorius and to articulate with the tympanic. The sclerotic plates preserved in the orbit of the subject of fig. 2 bend more abruptly than usual towards the back part of the cavity, suggesting a depressed spheroid form of the eyeball. The nasal (15) presents the usual connections; the margin which it contributes to the upper part of the nostril (re) is slightly convex, encroaching on that opening, which seems to be rather longer, proportionately, than usual : in the subject of fig. 2 it is 2^ inches in length. In the composition of the lower jaw, I have noted that the angular element extends a short way in advance of the surangular before disappearing externally. The point of the surangular enters a notch in the hind part of the dentary, about half an inch anterior to a line dropped from the fore part of the nostril. The angular disappears, as usual, between the splenial and dentary, not between the splenial and surangular. The splenials contribute a small proportion to the mandibular symphysis (Tab. XXI, fig. 4, 32')- The teeth (Tab. XX, fig. S) conform in relative size and slenderness of crown with the slenderness of the bones wielding them. Those at the fore half of the jaws inclinr more backward than usual, and hardly assume a vertical position in the maxillary ana post-mandibular regions. I have counted from sixty-five to seventy on each side of the LIASSIC ICHTHYOSAURS. 81 upper jaw, of which twenty-five, or thereabouts, are implanted in the maxillary. In the lower jaw there are about sixty teeth in each dentary (Tab. XXVIII, fig. 2, 33). A few detached teeth in a portion of a large Icli, tenuirostris from the Lias at Pyx Hill measure, each, 1 inch 4 lines, the enamelled crown being about a third of that length ; the cement- clad root is 4 lines in diameter, rather thicker in proportion than in the smaller-sized specimens of the present species. The vertebral column (Tab. XXVIII, fig. 1) agrees in general length with the charac- teristic shape of the head. In the best preserved specimens it is nearly four times the length of the skull. I have counted 156 vertebrae in a well-preserved column of the large specimen from Pyx Hill ; in this the pinnigerous part of the tail was two feet in length, and had been bent down iu the burial and subsequent petrifaction of the Sea-dragon at almost a right angle to the trunk; this deflected part included sixty centrums, which seemed to be relatively somewhat shorter as well as narrower than those of the trunk. At the bent part of the column the margins of the terminal articular facets were slightly deflected, and markedly raised from the level of the sides of the centrum, indicative of the degree and frequency of flexure at this part. The fore-and-aft diameter of a post-abdominal vertebra in an average-sized Tenuirostral is 13 lines, the vertical diameter being 2 inches 6 lines. The terminal articular surfaces of the centrums are more unifomily concave than in the previously described species. I have not found in Ich. tenuirostris more than two hypapophyses at the fore part of the column, one wedged between the basioccipital and the atlas, the other between the atlas and axis. This more simple apparatus for fixing the immediate support of the skull suggests an accordance with the lighter and more slender character of that part. The centrums gradually increase in fore-and-aft dimen- sions to the pelvic region, and do not begin to decrease in size till about ten vertebrae beyond the pai-t forming the base of the long caudal region. The ribs soon become long and slender as they recede from the head, and increase in length to near the hind end of the abdomen ; thence they shorten less gradually than usual. Forty-five pairs of the long and regularly-curved ribs show the external longi- tudinal groave. The parial fins (Tab. XXVIII, figs. 4, 6) show a somewhat less disparity in the size of the pectorals and ventrals than obtains in Ich. communis and Ich. intermedins. Their framework has fewer and larger bones, and the fore paddle impresses one with its massive proportions compared with the vertebrae. The clavicles are relatively more slender than in Ich. communis, but of the usual form, diminishing at the two extremities. The scapula is relatively larger than in Ich. intermedius, and is thicker and more expanded at the humeral end ; its fore border is moderately concave and longer than the hind one. The coracoid (ib., fig. 3, 52) has a broad neck supporting a large and thick scapulo-humeral articulation ; it has a deep and narrow anterior notch, and a shallow posterior emargi- nation. In a well-preserved specimen of the present species, in the Philosophical 11 82 BRITISH FOSSIL REPTILES. Institution, Birmingham, the length of the coracoid is 4 inches 5 lines, and the breadth 3 inches ; the length of the humerus of the same specimen is 3 inches 10 lines, the breadth of its distal end is 3 inches. The transverse diameter of the radius equals the antero-posterior diameter of the centrums of two of the parallel vertebrae ; its anterior margin is notched. The ulna has a corresponding size, with a smaller anterior notch circumscribing with an apposed notch in the radius a roundish vacuity. These bones were anchylosed together and to the humerus in the Birmingham specimen. The ' manus ' commences by three transversely oval carpals, of which the radial one is notched, as in the radius ; but this character is not repeated, as in Ich. acutirostris and Ich. platyodon, in the next distal bone, nor is the radial digit bifurcate, as in Ich. com- munis and Ich. intermedius. There are but three series of digital bones, with a fourth shorter marginal series of smaller ossicles. In the hind paddle (Tab. XXVIII, fig. 6) the femur, like the humerus, has a longer shaft than usual, and not so proportionally broad a distal end. The tibia is notched anteriorly like the radius, but not so deeply ; the corresponding tarsal bone is more feebly emarginate. In this fin, also, there are but three series of digital ossicles. In the Museum of the British Institution there is a skeleton of Ich. fenuirostris, tliirteen feet in length : it is from the Lias of Lyme Regis, Dorsetshire. Evidences of the same species have also been obtained from the Lias of Stratford-on-Avon, of Bristol, of Street, Somersetshire, and at Barrow-on-Soar, Leicestershire. /;. Ichthyosaurus longirostris, Ow. Tab. XXI, fig. 2; Tab. XXIV, fig. 3; Tab. XXVIII, figs. 7, 8, 9. The specimens in the British Museum, from the Lias of Barrow-on-Soar, on which the present species is founded, and the least incomplete of which is the subject of figure 7, Tab. XXVIII, have borne the above specific name in the Public Gallery of the Depart- ment of Geology for fifteen years. The description was reserved for the present Work, and I regret the unavoidable circumstances which have delayed its publication. Reckoning the lost terminal caudal vertebrae according to the guiding analogies I estimate the total length of the animal to include four lengths of the skull. Of this part the length of the rostral extension anterior to the orbit is four and a quarter times the antero-posterior diameter of that cavity : and yet the orbit is relatively larger than in Ichthyosaurus fenuirostris. The teeth correspond in size with the slenderness of their supporting jaws ; they are of difficult detection ; the best preserved show crowns, as in fig. 9, Tab. XXVIII. With sufficient magnifying power traces of longitudinal striae are discernible on the enamelled crown ; the cemented base is as little tumid as in Ich. tenuirostris. In the relative LIASSIC ICHTHYOSAURS. 83 minuteness of the teeth of Ich. hngirostris we may discern the transitional step to the edentulous Ichthyosaurs described by Prof. Marsh.^ I conclude that the present, together with other long- and slender-jawed Fish-lizards, may have preyed in a great degree upon the contemporary Cephalopods with internal rudimental shells as well as on Fishes. The eye in Ich. longirostris is proportionately large, like the orbit which the sclerotic circle almost occupies, suggestive of the nocturnal habits of the species ; the plates seem to have been not fewer than sixteen in number. The present s|)ecies, like Ich. tenuirostris, is characterised by the large size of the pectoral fin, and that of the ossicles representing the carpals and phalanges of its digits. These, however, are limited in number, as in Ich. tenuirostris; the three normal digits (ii, III, iv) are instructively conserved, though not terminally entire, in fig. 7, Tab. XXVIII. The supplementary ossicles here preserved are, situated as in the tenuirostral species, along the ulnar margin, forming a sort of rudiment of the digit (v), which is normally developed in Ich. communis and Ich. intermedins. The scapula, clavicle, and coracoid of the same side as the fin are definite, but dislocated. The coracoid repeats the type of that of Ich. tenuirostris so far as having the anterior notch the best marked, but the posterior one is more faintly indicated. The anterior notch (Tab. XXVIII, fig. 9) is placed further back, and becomes lateral rather than anterior. The articular prominence is also nearer the hind border of the lamelliform bone. The radius and radio-carpal bone are notched anteriorly, as in Ich. tenuirostris. The pelvic paddle has the same relative size as in that species, and the same tridactyle structure. I reckon forty-eight vertebrae between the skull and pelvis ; fifty-two vertebrae can be made out in the extent of the caudal region preserved ; to which may be added about a score more for the wanting pinnigerous terminal portion of the tail, which probably has been torn off" by predatory assailants of the dead reptile. Besides the Liassic locality of Barrow-on-Soar, specimens and parts of Ichthyosaurus longirostris occur in the zone characterised by ulSgoceras angulutum ; also in the zone of Arietites Bucklandi? I. Ichthyosaurus latimanus, Ow.^ Tab. XXIX, figs. 2, 7. This species is nearly allied to Ich. communis in the size and construction of the 1 " A New Order of Extinct Reptiles (Sauranodonta)," ' Amer. Journ. of Science and Arts,' vol. xvii, Jan., 1879, p. 85. 2 See the classical Monograph, by Dr. Wright, F.R.S., &c., in the volumes of thePalseontographical Society issued in 1878 and 1879. 3 'Report,' ut supra (1839), p. 123. 84 BRITISH FOSSIL REPTILES. pectoral paddle ; but the still smaller proportion of the pelvic one suggested the nomen triviale, which is vindicated, also, as a sign of specific difference, by the proportionally sliorter and thicker jaws and by the modification of the vertebral centrums (Tab. XXIX, tigs. 2 and 7) ; in this character the present species differs from Ich. hreviceps. Ich. latimanus resembles Ich. communis in the ventricose, subobtuse character of the teeth, or, at least, some of the more worn ones, of the twenty-nine which may be counted on each side of both jaws. The articular surfaces of the centrum are concave at the middle third, the rest of the surface to the circumference is flat. In the specimen of Ich. latimanus, 6 feet 10 inches in length, and in that of an Ich. communis, 5 feet 2 inches in length, the following were the respective dimensions of bones of the scapular arch and its appendage : Scapula, length of . . Coracoid, antero-posterior diameter Coracoid, transverse diameter Antibrachials, breadth of . Length of fore fin, humerus inclusive Breadth of ditto leh. latimanus. Ich. communig In. Lines. In. Lines. 3 4 . . 3 0 3 8 . . 2 4 3 2 . . 2 0 2 5 . 1 7 7 6 . 5 0 3 6 . . 2 4 The clavicle is proportionally thicker than in Ich. communis ; in the skeleton above cited it is 6 inches 8 lines in length. The head is relatively shorter. In the specimen of which the dimensions are above given the mandible is 1 foot 4 inches in length, while in the Ichthyosaurus communis above compared it is 1 foot 5 inches in length. In a specimen of Ichthyosaurus latimanus in the Museum of the Philosophical Institu- tion at Bristol I counted 114 vertebrae ; the terminal caudals showed in a greater degree than usual the compressed character indicative of the vertical tegumentary fin. Parts of the carbonised integument are preserved on the slab of Lias on which lies the above skeleton ; faint traces of integument lie above and beneath the deflected caudal vertebrae ; a broad patch remains about four inches beyond the last preserved centrum, though not the last of the series. This is the sole direct evidence I have as yet detected of the tegumentary part of the tail-fin. Traces of the abdominal integument appear to be smoother than in the similarly preserved skin figured in Buckland's, ' Bridgewater Treatise.' If, as has been suggested (p. 44), the pectoral arch and fin relate to occasional reptation on the sea-shore, it may be inferred, from the partial flattening of the articular surfaces of the vertebral centrums, as well as from their proportions, in the present species, characterised as it, also, is by more massive proportions of the pectoral arch and greater LIASSIC ICHTHYOSAURS. 85 relative size and strength of the fore paddles, that it was more littoral in its habits than the majority of these marine Saurians. Saltford, near Bath, and the Penarth Beds (Rhcetic) of Glamorganshire are among the localities of Ichthyosaurus latimanus. m . Ichthyosaurus brachyspondylus, Oto. Tab. XXIX, figs. 3 — 6. This species is founded on vertebral characters, the centrums being shorter in proportion to their height and breadth than in any other that has come under my observation. In the abdominal centrum, the subject of figures 3 — 5, Tab. XXIX, the breadth of the articular terminal surface (ib., fig. 4) is 2 inches 11 lines; the vertical diameter is 3 inches, while the antero-posterior extent does not exceed 11 lines. In a more posterior centrum (ib., fig. 6) in which the diapophysis (d) has descended to contact with the parapophysis, the same proportions are preserved with slight diminution of size. In a larger raid-dorsal centrum of this species which I examined in the private collection of Mr. Rose, of Swaffham, the breadth was 3 inches 8 lines, the height 3 inches 9 lines, but the length did not exceed 1 inch 5 lines. In a collection of fossils from Mid-Jurassic beds in the Province of Moscow, sub- mitted to me by Col. Kiprianoff, in 1853, were centrums showing the same dimensional characters, together with a low medial ridge dividing the under surface, which is, like- wise, present in the British specimens. Col. Kiprianoff adopted the name, with my determination, of his Ichthyosaurian specimens, other evidences of which have since been detected and described in a valuable contribution to Russian palaeontology by Prof. Trautschold.i D. Conclusion. Although a study of the evidences of Ichthyopterygian organisation, of which, and its modifications, as exemplified at the Liassic period, the results are given in the foregoing pages, has left the impression, mainly, of the great additions which wait to reward subse- quent cultivators of this field of comparative osteology, I am unwilling to quit it without giving expression to some of the general notions which its cultivation has suggested. Palaeontology has been regarded, if not defined, as including a kind of knowledge of parts, or of structures, in such degree interdependent that, any one being given, others 1 Trautschold (Prof. H.), ' Erganzung zur Fauna des Eussiscben Jura,' 8vo, 1876, p, 5. 86 BRITISH FOSSIL REPTILES. may be deduced as a necessary consequence, such deductions being determinative of the relations of the whole ; so far as to give the Knower a power of predicating results, both zoologically, as respects the affinity of the otherwise unknown whole, of which only a part affects the senses, and physiologically, as respects the living powers of the whole and the part such extinct organism played in the sphere of its existence. This necessary connection and interdependency of the links of structures consti- tute the essential condition and attractive character of palaeontological science. The subjects, nevertheless, of the present Chapter, constrain me to submit the question, how far this science of ours has advanced towards sustaining the foregoing definition ; in other words, to how much of organic Nature at large, or of particular organisms, it is so applicable ? Let us suppose, for example, that no other part of the petrified frame of an Ichthyo- saur had come to our hands than had reached those of Scheuchzer or of Bronn, — a few vertebral centrums, for example, from the hind part of the trunk. Could we have otherwise concluded than they did ? Certainly not, had our knowledge of the vertebral structures been restricted to the same parts of the extinct and fossilised animal. Biconcavity of centrums is a pre-eminently piscine character, but not without excep- tions in the class of Fishes, even in that great proportion of the species whose osseous development has advanced to individualised bony segments of the spinal column. Such an exception, e. g. we have in the opisthocoelian vertebrae of the Bony Gar {Lepidosteus). But no known kind of Fish possesses vertebral centrums with both upper and lower transverse processes (' diapophyses ' and 'parapophyses '). The presence of these in certain of the biconcave vertebrae of Icihyosaurus, bespeaks that of ribs having a two-fold articulation with their vertebrae ; and such structure of rib implies a body-cage adapted to the movements of expansion and contraction of its cavity, which cavity we infer, there- fore, to have contained bags receiving air, and to have had associated movements for the purposes of respiration. But this function raises the exerciser above the grade of the Fish, and the next ques- tion of the Palaeontologist would be, whether the air-breather was cold-blooded or warm-blooded ? The biconcavity of the vertebrae would sustain the first conclusion, and consequently a reference of the extinct animal, so fragmentarily indicated, to the Reptilian, not the Mammalian, class. But what further insight into the nature of such Reptile could be gained by contem- plation of a solitary centrum, or even of a series of vertebrae ? With me no further step could be taken toward a sure knowledge of the nature of the cold-blooded air-breather, so partially indicated. A suspicion, at most, of an aquatic medium, and conseqviently of limb-structures adapted to locomotion therein, might have crossed the mind. But a complete reconstruction of the extinct animal, or certain knowledge of such, could only be the result of acquisition and comparison of the anatomy of the cranium, as well as of LIASSIC ICHTHYOSAURS. 87 the limbs and their sustaining arches ; and such has been the knowledge supplied by the sulijects of the foregoing pages of the present Chapter. Here I may remark that instances of ancient extinct forms, manifesting a more gene- ralised type, are more than ever worthy of note in the present phase of biological science ; and the Iclithyopterygia contribute a welcome addition to this suggestive class of pheno- mena. In the construction of their chief natatory organ for forward movement may be discerned a combination of mammalian, saurian, and ichthyic conditions. In the great length and gradual diminution of the caudal series of vertebrae may be noted the saurian character ; the tegumentary expansion, unsupported by bony rays, recalls the main feature of the cetaceous tail-fin, while its vertical position in the air-breathing Saurian brings it in close parallel relation with the corresponding natatory propeller in the class of Fishes. Nevertheless, the Ichthyosaur, as an aquatic air-breather, might be supposed to have exchanged, at a loss, the disposition of its terminal fin in comparison with its aquatic warm-blooded, fish-like successors ; but the pair of pelvic fins, wanting in all Cetaceans, are superadded to the locomotive instruments in the Ichthyosaurs, and were, doubtless, actively applied to bring the nostrils, when needed, within range of the superaqueous atmosphere. But in almost every extinct natural group of animals peculiar conditions present themselves. In no known cold-blooded Fishes was the visual organ so well, or so con- spicuously adapted to the detection of the finny prey as in ovir present subjects. To unusual size of eyeball, which in Dr. Buckland's experience sometimes reached that of a man's head, was added a circle of concomitantly large bones — the ' sclerotic plates ' — of form and structure in harmony with the requirements of the visual outlooks. Was a near object to be detected, the retraction of the bony circle and contraction of its aperture, surrounded by the laterally overlapping plates, would coincide with a concomitant con- vexity of the cornea pressed upon by the sqiieezed humours within and with the contrac- tion of the pupil — conditions concurring in the needed microscopic application of the eye. If the distant expanse of waters was to be scanned, the resumption of the normal position of the sclerotic ring and of the relaxed relations of the overlapping plates would coincide with an expansion of the pupil and flattening of the cornea, whereby the eye would thus acquire a telescopic range. Moreover, the wide transparent corneal aperture of the con- spicuously large organ of vision would enable the predatory Saurian to take advatage of the least amount of light penetrating the depths of its marine environment. THE FOSSIL REPTILIA OF THE LIASSIC FORMATIONS. CHAPTER III. Oedee— DINOSAURIA, Owen. Genus — Scelidosaueus, Owen. In the year 1858 a few fragmentary fossils of limb-bones were submitted to my inspection by James Harrison, Esq., of Charmouth, Dorsetshire, obtained from the upper part of the " lower Lias," near that place. They included portions of a femur and of a tibia, in which the texture of the wall and the size of the cavity of the shaft showed them to have been parts of a Saurian of more terrestrial habits than any of those which had been previously discovered in those liassic deposits : traces, moreover, of the extent and direction of certain processes, although broken away in the fossils, were discernible, which led me to suspect they belonged to a reptile alUed to Iguanodon. I therefore briefly notified the fact of a liassic Dinosaur in my ' Palaeontology,' ^ and indicated the animal by the generic name SceUdosaurus.^ The femur of the Iguanodon, as shown in vol i, p. 310, vol. ii, PI. 20, is characterised by the deep and narrow fissure dividing a compressed external trochanter from the head of the bone, and by a process from the middle of the shaft, on the inner side, opposite to the part where the " third trochanter " projects in some of the large herbivorous mammals {Perissodadi/Ia). Both these characters were repeated in the specimen of the shaft of the femur first submitted to me ; but the shaft, viewed sideways, showed a more decided sigmoid flexure than in the Iguanodon, and the fissure between the great trochanter and the 1 8vo. ed., 1860, p. 258. 2 Gr., (TKeXis, limb, aavpns, lizard; from the indications of greater power in the hind legs than in most Saurians, 12 90 BRITISH FOSSIL REPTILES. proximal end of the bone was relatively deeper. This end, divided by the cleft from the great trochanter, was subcompressed from side to side below the swelling out of the head, which had been broken or abraded away, showing a fine cancellous structure at that part. The antero-posterior diameter of this part is 6 inches ; the transverse diameter, opposite the base of the outer trochanter, is 3 inches 8 hnes. The fore part of the shaft showed at its upper half a flattened, oblong, rather rough surface for muscular implantation. Below, and on the outer side of this surface, was a rough, roundish, slightly prominent tuberosity (s), continued at its inner side into a ridge, which descends with a slight curve outwards on the fore part of the middle of the shaft of the femur, where it terminates in a point at v. These risings indicated the force of the large muscles acting upon the limb, and by their insertions raising and drawing forward the femur. Behind the base of the process was a large, oblong, rough ridge indicating the extension of the surface of attachment, behind and beyond the process itself, for a powerful muscle depressing and drawing back the femur. From the great trochanter a narrow, rough surface, not projecting as a ridge, extended nearly straight down the outer and back part of the shaft. Exterior to this surface was an oval foramen, most probably for the passage of the blood-vessels and nerve to the medullary cavity. The transverse section of the middle of the shaft is nearly circular ; the thickness of the compact wall of the medullary cavity is here about one sixth of the transverse diameter of the bone. I have not seen a bone of any other Dinosaur indicative of more vigorous action of the hind limbs than the present femoral shaft. The foregoing instructive fossil was accompanied by the shaft of a til:)ia of corresponding size, crushed and broken at both ends ; it measured 18 inches in length and 2 inches 8 Unes in diameter at its middle, the circumference of the shaft there being 10 inches. These proportions indicated a hind leg, longer and more slender, relatively to the trunk, than in the Megalosaur, Iguanodon, or other Dinosaur with which such comparison may be made. The bone being fractured across the middle of the shaft, shows a large medullary cavity ; the compact, bony wall does not exceed 3 lines in thickness, the cavity itself being 1 inch 3 lines in diameter. At the proximal end the antero-posterior expansion and its ridges have been broken away. The bone gradually contracts, as it descends, to a subtriedral shaft, with a triangular transverse section, two of the angles being rounded off, and the third remaining, which was opposite the fibula. The distal expansion has been, in like manner, broken away ; but its commencement shows the rise of an anterior ridge in addition to the fibular one. I shortly after received from LIASSIC DINOSAURS. 91 Mr. Harrison the lower half of a right femur and the upper half of the right tibia and fibula, cemented by the matrix in the natural relative position in which they enter into the formation of the knee-joint, when bent. This remai-kable specimen indicates the tranquil state of the sea-bed or bottom after it had received the dead carcass of the Dinosaur. No agitation or other external violence had displaced the bones of the leg after the solution of the ligaments which tied them together in the living animal ; when the depth to which they had sunk, and the consistency of the mud or clay bed, tended to retain them in their natural position. The portion of femur preserved indicates a slight backward bend of the shaft, which at the fractured part — probably a little below the middle of the bone — presents an almost circular transverse section. The circumference here is 10 inches ; the compact wall of the bone is 6 lines thick ; the medullary cavity 2 inches in diameter. A little below the fractured end, and 8 inches above the lower end, the shaft shows the termination of the characteristic inner process. From this point the femur expands gradually, and chiefly in the transverse direction. Posterioi'ly it becomes impressed by the popliteal cavity, which deepens and widens to the upper and back part of the inner condyle ; which, by its production towards the outer condyle, contracts the lower end of the popliteal cavity transversely. On the outer side of the distal expansion of the femur the external wall is in part broken away ; but a shallow and narrow longitudinal impression is indicated, terminating below in a rather shallow notch, which marks out the inner and hinder part of the outer condyle from the outer part of the same condyle. This notch corresponds with that between the tibia and fibula, and defines the portions of the outer condyle assigned to those bones respectively. The inner condyle is rather flattened on the inner side. The tibia is much expanded at the proximal end, chiefly by an extension of the bone forward; it is slightly convex on the inner or tibial side; a longitudinal prominence extends from the fibular side of the expansion, near the fore ipart, answering to the ectocnemial process in the bird's tibia ; the main expansion forms the procnemial process which has subsided to the ordinary level of the shaft about six inches down the bone. The back part of the proximal end of the tibia presents two almost hemispheric protuberances, side by side ; they might be mistaken in a detached bone for the backwardly projecting condyles of a femur, but are less deeply severed. The outer tuberosity articulates with a slight depression in the contiguous part of the fibula. The fore part of the proximal portion of the tibia is, transversely, concave, exterior to the pro- and ectocnemial processes. The fractured part of the shaft, eleven inches below the knee-joint, presents a full, oval section, with the same proportion of compact bony waU to medullary cavity as in the femur ; the white spar filling the cavity contrasts strongly with the jet-black colour of the petrified bone. The 92 BRITISH FOSSIL REPTILES. transverse diameter of this part of the shaft is 2 inches 3 lines ; the fore-and- aft diameter is 2 inches 6 lines. The fibula expands chiefly in the fore-and-aft direction at its upper end, where it measures 5 inches across. Six inches lower down this diameter has contracted to one of 1 inch 8 lines ; eleven inches lower it measures 1 inch 3 lines, the transverse diameter being 9 lines. Seven inches from the proximal end the fibula presents at its outer and back part a thick, longitudinal, rough ridge, for the attachment of a muscle. It continues in contact with, and gets rather behind, the tibia as it descends. The foregoing indications of a Dinosaur in the lower Lias excited speculation as to whether it had been herbivorous, like the Iguanodou of the newer Mezozoic beds, or carnivorous, like the Megalosaur, which has been traced from Wealden down to the Great Oolite. The structure of the femur jjointed the former way, but the proof which the dentition only could give was wanting. The persevering encouragement afforded by Mr. Harrison to the workmen in the Lias quarries was subsequently rewarded by the acquisition of the fine specimen of a skull which forms the subject of Plates 45, 46, 47. The teeth, in their close-set, thecodont implantation, relative size to the jaw, degree of expansion, and general shape of the crown, resemble those ascribed to the Eijlceosaurus (Vol. i, p. 367, PI. 39) ; but the crown presents the median longitudinal prominence and marginal serrations which bring it closer to the Iguanodont pattern ; and, in the degree in which they depart therefrom, they still more closely resemble the teeth of the Echinodon from the Purbeck, which may prove to be a small kind, or young, of a Dinosaur. They, however, present diS'erent proportions. Referring, therefore, the skull in question to the Dinosaurian order, it supplies most acceptable information as to the cranial structure of that group, in addition to that derived from the Iguanodon Foxii of the Wealden beds (Vol. i, p. 520, pi. 49, figs. 9, 10). Of Megalosaurus, and Hylceosaurus, portions of lower jaw, and fragments of the upper jaw, palate, and basis cranii, are all that have hitherto come to light. But the present specimen is the entire skull, wanting only the fore end of the upper and lower jaws. The cranium has been slightly crushed and distorted by oblique pressure, due to movements of the matrix after imbedding and petrifaction. The right halves of the mid-frontal and nasal are depressed a little below the level of the left halves of the same bones, and the right diverging branch of the parietal has been broken from the rest of the bone, near the median line, and dislocated by the same pressure from its union with the mastoid. The right ramus of the LIASSIC DINOSAURS. 93 mandible, accompanying the movement of that side of the head, has been pushed so far below the left ramus as to have its inner side brought into view below that of the left side of the skull. The occipital conforms to the Lacertian type (PI. 60, fig. 2, 4), in the pro- portions and direction of the par-occipital ; this process is long, narrow, straight, directed outwards, compressed from before backward, and slightly expanded at the extremity, which is applied to the back part of the mastoid and tympanic at the junction of those bones. It has been slightly displaced, its end appearing on the left side at 4, PI. 45, with matrix intervening between it and the tympanic (28). A part of the exoccipital which projects backward to contribute to the formation of the condyle is exposed near the mass of matrix, including the atlas vertebra and nuchal dermal bones. The cranial part of the skull, posterior to the orbits, is shorter in proportion than in the lizards, and resembles, in this respect, that of the Iguanodon (PI. 49, fig. 9) and crocodiles. The parietal is short, and bifurcate behind, as in lizards. The body of the bone, or part between the temporal foss«, is subcom- pressed where it forms the smooth, concave, inner sides of those depressions, which do not meet above, but are separated by a narrow, flat tract ; this might be converted into a ridge in older individuals. The fore part of the parietal slightly expands where it is overlapped by the frontals. Bach hind branch of the parietal extends outward and a little backward ; its pointed end is obliquely overlapped anteriorly by the mesial branch of the mastoid, completing therewith the hind boundary of the temporal fossa. The crushed and dislocated state of the calvarium along its middle line does not permit the usual evi^lence of a foramen parietale to be detected, but the appearances are against such perforation being present. This foramen is not constant in modern lizards ; the Scelidosaurus may agree with Cydodus and Tejus in this respect. The parietal bone, as a whole, plainly accords with the lacertian, not with the crocodilian, type of that bone. The mastoid (8) is a triradiate bone, forming the upper and hinder angle of the cranium, from which one ray passes mesiad to join the parietal (7), a second ray forward to join the post-frontal (PI. 46, 12), and a third ray downward (PI. 45, 8), to join the tympanic (2s). A fracture of the body of the mastoid, by which he anterior branch is broken away on the left side, exposes a cancellous cavity, probably forming part of the organ of hearing. The two halves of the mid-frontal (PI. 47, ll) have been separated along the medial line, and the right half depressed. The separation appears to have been at a suture, as is certainly the case with the nasal bones ; the medial margin of three fourths of the left frontal show the jagged, sutural character. I conclude, therefore, that the mid-frontal was divided, as in Iguanodon (PI. 49, fig. 9, 11), and as in Varanus and Lacerta proper ; and that it was not a single bone, as in 94 BRITISH FOSSIL REPTILES. the Iguana and most Lacertilia, and as it is in the Grocodilia. Each half of the frontal in Scelidosaurus is a long, inequilateral triangle, the medial being the longest side, the posterior, which joins the parietal, the shortest ; the antero- external border is irregularly and deeply notched, uniting with the post-frontal, super-orbital, pre-frontal, and nasal bones ; it is excluded, as in Lacerta proper, by the large super-orbital bone (71) from the orbit. The outer surface of the frontal is sculptured by irregular lines and grooves, but less deeply than in Crocodilus. The post-frontal (12) forms the back and part of the upper border of the orbit, uniting with the super-orbital, the frontal, and malar, and sending backward an angular process to join the mastoid, completing the upper bar or zygomatic arch of the temporal fossa. This arch had been broken away on the left side (PI. 45), but is preserved on the right side (PL 46, 8, 12). The pre-frontal (Pis. 45, 46, 14) presents a horizontal and a vertical portion ; the former and larger part is wedged between the frontal, super-orbital, and nasal bones, the descending plate joins the lacrymal (73), and touches the upper angle of the maxillary (21). In the Crocodile the aspect of the whole outer plate of the pre-frontal is upward ; in some Lacertians the major part looks outward. The nasal bones (15, Pis. 46, 47) unite above and behind with the frontal (11) by a short border, obliquely and irregularly cut, to include the pointed anterior ends of the lateral halves of the frontal; the nasals expand as they advance, in union, first, with the pre-frontals, then with the maxillaries, where they slightly decrease in breadth. The outer plate of the nasals looks upward; the maxillary border is slightly bent down (15, fig. 2, PL 46), and is overlapped by the maxillary (21, ib.). The mutilated fore part of the skull precludes the determination of the relations of the nasals with the pre-maxillary, and of the character of that bone ; but it most probably repeated, in the main, the conditions which it presents in Iguanodon.^ The fracture shows the superior thickness of the median and lateral borders of the nasals, the intervening part being, as it were, channeled below for the air- passage ; this has not here been divided by any ossified vertical septum ; the thickened palatal and alveolar parts of the maxillary, as they bend toward each other, present a convexity transversely to the nasal passage. This is closed below, as it seems, by the vomer (PI. 46, fig. 2, 13). Of the hind part of the bony palate the pteiygoid was brought into view by removing the matrix between the diverging rami of the mandible. The body of the bone is iu the form of a subtriangular plate, of 1 inch 7 lines extent along its mesial border, which is slightly concave, receding from its fellow at the medial line, or base, as in the Iguana ; the apex extends outward, and a little downward to abut against the fore and inner part of the ectopterygoid. From the hind 1 Vol. i, p. 521, pi. 49, figa. 9, 22. LIASSIC DINOSAURS. 95 border near the base a long and narrow process is sent off to abut against the tympanic. There is no trace of teeth on the pterygoid, as in the recent Iguanas ; the higher type of Sam-ian dentition is retained in Scelidosauriis as in Iguanodon (PI. 60, fig. 5, 20, 24). The hind and probably main part of the maxillary, here preserved, is chiefly remarkable for the horizontal ridge which nearly equally divides the outer or facial plate of the bone into an upper and lower facet ; and this ridge is continued a little way below the orbit upon the malar bone. It corresponds with the more strongly marked ridge in Pti/chognathus and Oudenodon} There is a lower and slighter longitudinal prominence of the maxillary along the outer alveolar plate. The maxillary reaches back beyond the middle of the orbit, from which it is separated, as in other Saurians, by the malar and lacrymal bones. On both sides there is a small, unossified space between the maxillary and lacrymal ; this corresponds with the larger vacuity in that part of the bones of the face in the Pterodactyle, which is reduced to the present proportions in some Teleosaurs, and becomes the functional nostril in the Ichthyosaur ; but I beheve that the true external nostrils of Scelidosaurus were included in the fore part of the skull which has been broken away, and were, as in the Teleosaiir, distinct from the maxillo-lacrymal vacuities. The orbits of Scelidosaurus are subcircular, almost vertical, looking outward. Were the super-orbital ossicle in Grocodilia enlarged and fixed by suture in the upper scoop of the orbit, it would give a less vertical outlook to the eye than it usually presents, especially in the skull of a crocodile from which that ossicle has been removed. But the composition of the rim of the orbit in Scelidosaurus is open to other homologies. The bone (71) may be compared with that wedged into the upper and back part of the orbit in some lizards, between the frontal and post-frontal, and by Cuvier regarded as a dismemberment of the latter element ; only in Scelidosaurus it is extended forward to the pre-frontal, excluding the frontal from the orbit. In Ichthyosaurus the post-frontal has a forward extension to junction with the pre-frontal ; it also passes backward to join the mastoid, leaving to the bone at the back of the orbit (PI. 20, fig. 1, 12) a simple post-orbital function. In Scelidosaurus the bone which joins the mastoid sends down a post-orbital bar (PI. 40, 12) to join the malar (ib. 20). The post- frontal holds the same relations to the orbit in Iguanodon (vol. i, p 521, PL 49, fig. 9, 12). In both genera there is a masto-post-frontal zygoma, as well as the ordinary malo-squamosal one. But the intervening space is not walled over by a supplementary plate as in Ichthyosaurus (PI. 20, fig. 1, 27^) The delicate lacrymal bone (Pis. 45, 46, 73) appears to have been 1 ' Description of the Fossil Eemaiua of the Eeptilia of South Africa,' 4th, 1876, vol. i, pp. 48 — 56 ; vol. ii, pis. xlv, Iv. 96 BEITISH FOSSIL REPTILES. fractured on the left side ; on the right side it is entire. The malar bone (26) begins anteriorly, in a pointed form, between the lacrymal and maxillary, increases in depth as it extends beneath the orbit, sends up a process which bifurcates to receive the point of the post-frontal in the cleft, and extends backward and downward as a slightly convex and somewhat roughened plate, which articulates by its lower convex, but somewhat irregular, border with the squamosal (27). The posterior border of the malar presents a regular and well- defined, concave curve. The chief peculiarity of the bone is its unusual vertical extent posteriorly. The squamosal (Pis. 45, 46, 27) articulates with the lower border of the malar, and expands to be articulated with the outer part of the lower half of the tympanic (ib. 28). This deep and powerful arch of bone, answering to the zygoma in mammals, afforded attachment to large, masseteric muscles operating upon the lower jaw. Similar muscles may have been extended between the ridges of the upper and lower jaws. The tympanic (partly exposed in PI. 45, 28) is a long bone, compressed from before backward, almost vertical in position, with a slight forward bend, but firmly wedged between the mastoid and par-occipital above and between the squamosal and pterygoid below. The back part of the tympanic is convex transversely at its inner half, concave at its outer half, where the margin is slightly produced to join the upper part of the squamosal ; the inner part of the tympanic is more extended where it is overlapped or abutted on by the pterygoid. Below this expansion the tympanic becomes contracted and thickened, forming a kind of neck to the transversely extended convex terminal condyle. In the vertical position and length of the tympanic, Scelidosa^lrus resembles the Lacertia ; in its fixity and extent of its connections it resembles the Grocodilia. The lower jaw includes in each ramus an articular (29), a surangular (30), a coronoid (30'), an angular (31), a splenial (32), and a dentary (33) piece. The articular (PI. 47, fig. 2, 29) is situated in the inner side of the sur- ano-ular (PL 45, 3o), and is thickened and projects inward to form the cavity for the major part of the tympanic condyle, the outer border of which rests on the surangular. This element, convex externally, presents a longitudinal ridge near its upper part, which rises to join the posterior angle of the dentary element in forming a low coronoid process. The angular (ib. 31) does not extend beyond the surangular, but makes with it the angle of the lower jaw ; it grows in vertical extent as it advances, is convex externally, unites with the dentary, and sends forward from its lower part a pointed process between the dentary and splenial elements. The splenial (PL 46, 32) makes a small appearance on the outer side of the ramus, between the angular and dentary (33), but is chiefly visible as a broad, smooth plate (PI. 47, fig. 2, 32), applied to the inner side of the dentary. The LIASSIC DINOSAURS. 97 dentary (Pis. 45, 46, 33) is a very powerful bone, with the outer surface divided into an upper and lower facet by a longitudinal ridge paralleling that of the upper jaw. The ridge, commencing near the base of the coronoid process, descends, describing a slight curve to the middle of the outer surface of the dentary. Below the ridge the bone is convex, above it is concave ; the lower facet has the kind and degree of roughness observable on the exposed surface of most of the cranial bones ; the upper facet has a smoother surface, corresponding in that respect with the surface below the ridge of the maxillary. The foregoing character of the lower jaw has, hitherto, been observed only in a fossil one, which has been referred to the Dinosaurian order ; by Mantell,' originally to Igimnodon, and afterwards, when it had been shown to be more probably part of the Eylceosaurus,^ to a genus which he called Begnosaurus? In this specimen the outer surface of the dentary is divided into an upper and lower facet by a longitudinal ridge, which, commencing near the upper margin, probably at the base of a coronoid rising, descends as it advances to midway between the upper and lower border. It is, however, more obtuse than in Scelidosaurus, but the upper facet presents a Hke smoothness and vertical concavity.* In size the specimens closely correspond, and also in the close arrangement of the series of teeth. But these were relatively smaller and more numerous in the Wealden fossil ; for whereas in Hijlceosaurus ten teeth, or their sockets, occupy an extent of 1 inch 8 lines of the alveolar border, the same extent includes only seven and a half teeth or sockets in Scelidosaurus. In this genus, moreover, the ramus of the mandible presents a curve convex downwards, to about the same degree as the opposite curve is presented by the corresponding part of the jaw of Hiflceosaurus. in which this peculiar bend is noticed in Vol. i, p. 366. In the mandible of Scelidosaurus a ridge, corresponding, perhaps, to the lower ridge in Eylmosaunhs, is situated further back and higher up upon the surangular ; and the facet, concave vertically between the lower ridge and the beginning of the upper ridge, is peculiar to the mandibular fragment referred to the Hijlceosaurus. Thus, with corresponding Dinosaurian character, imparting robust strength to the mandible, there are well-marked generic distinctions in the specimens here compared, both in the conformation of the jaws and teeth. The mandibular rami of Scelidosaurus describe a slight sigmoid curve from the angle forward, horizontally, at first concave, then convex, towards the median line, where they meet without blending at the part fractured. It is not probable that the symphysis would be much prolonged beyond this point. The degree of convergence of the contour lines of the whole skull, both median and lateral, with 1 ' Wonders of Geology,' 1838, vol. i, p. 393. 2 " Eeport on British FoFsil Keptiles," 'Trans, of Brit. Association,' 1841, p. 120. 3 ' Philos. Trans.,' 1818. * Vol. ii, pi. 39, fig. 1. 13 98 BRITISH FOSSIL REPTILES. the decreasing size of tlie anterior teeth, makes it more probable that but a small proportion of the muzzle is wanting in the present specimen. Removal of matrix from the hinder interspace of the mandibular rami exposed a ceratohyal, 3 inches in length, 4 lines in breadth, of a slight sigmoid flexure, the hind part bending up to between the angle of the mandible and the atlas vertebra ; as no trace of basihyal was foimd, this element was probably cartilaginous, like that broad one in the tongue-skeleton of the crocodile.' The specimen of SceUdosaurus here described has been buried and petrified with the mouth shut ; there has been no dislocation of the under jaw, and the skull shows that the teeth of the upper jaw overlapped and concealed those of the lower. The crowns of both series were a little inclined inward, as shown at the fractured fore part (PI. 46, fig. 2, a b,) ; there is a similar inclination of the alveolar plates. The teeth are small, or of Lacertian proportions to the jaws ; they are numerous and close-set, implanted in sockets (ib. a) forming an uninterrupted series along the alveolar border. The faug is simple, and longer than the crown, presenting a full ellipse in transverse section, and projecting a little beyond the socket. In the upper jaw the crown (PI. 46, fig. 3, magn.,) begins by bulging outward, with a smooth convexity subsiding as it gradually expands, and dividing to be continued along the middle and the margins, with intervening concavities, producing an undulated surface across the broadest part of the crown. The marginal con- vexities or ridges terminate each in a point at the broadest part of the crown ; whence, the plate-shaped tooth having thinned off to an edge, this is divided on each side into five or six smaller points ; these denticulate margins converge straight, at an angle rather less than a right one, to the apex of the tooth, which is formed by the pointed termination of the median convexity. The crown is coated by a polished enamel, of jet blackness in the fossil, smooth under the lens upon the convexities, finely punctate in the hollows of the expanded part of the crown. The whole tooth in the upper jaw is very slightly bent backward, with as slight an oblique twist, making the hinder angle overlap, in some, the front angle of the crown of the tooth behind. The inner surface of the hind teeth exposed in the right ramus of the jaw (PI. 47, fig. 2) shows similar configuration of the crown to that of the outer surface of the teeth above ; but with a larger proportion of the serrated part, and with the borders less equal, the anterior one showing as many as nine points. On the left side, in an extent of the alveolar border of the upper jaw measuring 4 inches, there are nineteen sockets, and only one tooth missing. On the right side, in an extent of 3^ inches, there are sixteen sockets, and three teeth missing. The fractured part of the jaw yields evidence of the usual 1 ' Archetype of Vertebrate Skeleton,' 8vo, 18i8, p. 121, fig. 22, 40. LIASSIC DINOSAURS. 90 reptilian provision for successional teetli in reserve alveoli, containing tooth- germs, at the inner side of the base of the teeth in place (PI. 46, fig. 2, c). The teeth gradually increase in size from the hindmost to the fifth in advance, continue of about the same size to the tenth, and then gradually decrease in size to fractured fore part of the jaw. Were the serrated borders of the terminal half of the crown to be worn down, the teeth of Scelidosmtrus would be like those referred to Eylmosaurus in Vol. i, p. 377, Vol. ii, PI. 39. There is no evidence, however, that any of them have been so worn down ; in this respect they resemble more the teeth of Echinodon, the upper teeth in Scelldosaurus differing chiefly in the proportions of length to breadth of the crown. Whether the anterior teeth had the simple laniariform character at the fore part of the jaws in Scelldosaurus, as in Iguanodon and Echinodon, remains to be proved. The finely and sharply serrated and pointed teeth of the Scelidosaurus glided upon each other, the upper on the outer side of the under, like the blade-shaped crowns of the carnassials of feline mammals ; and yet the similarity of the teeth, in their number and uniformly small size, to those of the modern Iguanas suggests that they may have been put to like uses. The compressed, serrate crowns in those herbivorous lizards worked obliquely upon each other, in a similar scissor-blade way. In Iguanodon the dentition is obviously modified more decidedly for mastication of vegetable substances. In Scelidosaurm it is adapted for division of such substances, but it would be equally efi'ective in piercing and cutting or tearing through animal textures. If this Dinosaur occasionally went to sea in quest of food, it may be expected to present in the fore part of the jaws, wanting in the present specimen, laniariform teeth, as in Iguanodon (PI. 49, fig. 9, i), for the prehension and retention of living prey. Should these prove to be absent, and the dental series to begin as it ends, it will incline the balance of probability to the phytophao-ous nature of the Liassic Scelidosaurus. Following in the track opened out by the discovery of the skull, about twelve successive blocks of Lias were secured, with more or less evident indications of included bones, all of which, together with the skull, were secured and transmitted to the British Museum. Subsequent complete exposure of the included organic remains has brought to light the entire vertebral column of the trunk and tail, to very near the termination of the latter ; the scapulo-coracoid arch and part of one fore Hmb being associated with the thorax, and the iliac bones and both hind limbs with the sacrum : all were parts of the same individual Dinosaur. In the operation of clearing off the matrix, scattered dermal bones first presented themselves, and these were removed, with a note of their position, when it became plain that they did not touch or rest upon any part of the endo-skeleton. This 100 BRITISH FOSSIL REPTILES. being readied, the dermal bones in contact with it were left, save where tliey concealed some joint, process, or other light-giving or characteristic part of the framework. In the coiu"se of these operations it soon became evident that the whole vertebral column, in a series of consecutive and but slightly disturbed and mostly coarticulated segments, from the axis to the thirty-fourth caudal vertebrae inclusive, had been raised from their place of deposit ; all the parts, save the centrum and a small and low coalesced neural arch, having ceased to be developed, in the terminal caudal vertebrse, the last of whicb in the recovered series was reduced to dimensions so small as to indicate that but very few remained to complete the tail of the Scelidosaur. The first vertebra of the neck was adherent to the back part of the skull above described. Of the liassic masses following that which included the skull the first four contained twenty vertebrae, extending from the axis to the mass including the sacrum, and they were clearly consecutive save at one part of tbe neck. The back part of the mass containing the skull includes the atlas vertebra in connection with the occiput, and surmounted by a pair of dermal bones (PI. 48, fig. 1, (In). The block whicb fits to the fractured surface including the body and tbe neurapophyses of the atlas contains the axis and third cervical vertebra. The next piece revealed one neai'ly entire cervical vertebra (ib., figs. 3 and 4) and part of a second vertebra. The third, larger, piece included ten coarticulated vertebrae (Pis. 49, 50), but the continuity of the fore part of this mass with the last mentioned could not be clearly made out. The fourth block fits to that containing the ten dorsals, and included the five consecutive vertebrae with part of a sixth (PI. 51). The block which contains the sacrum has also two vertebrjB in advance of it (PI. 53), part of the first of which lies in the preceding block. Thus there was evidence of at least twenty-two "true " vertebrte; but there may have been one or two vertebrae from the region of the neck which had not been recovered. The vertebra attached to the first sacral seems not to have sup- ported ribs : the one in front of it has a pair of long, freely articulated ribs, and may be reckoned the last dorsal. Including this, there may be assigned sixteen vertebrae to the dorsal series, if we include therein the ten vertebrae in PL 49, leaving sLxor seven to the cervical series. The lumbar series seems thus reduced to one vertebra. The sacrum includes four vertebrae. Of the caudal series thirty-five vertebrae were preserved, in five consecutively fitting blocks of matrix, leaving parts of two terminal ones, so small and simple as to show that very few are wanting in the present fossil skeleton. The modifications of the spinal column of the trunk and tail of the Scelidosaur could thus be studied and compared in sixty consecutive vertebra of one and the same individual. The fracture of the matrix including the skull has passed thi'ough the ceu- LIASSIC DINOSAURS. 101 trum (PI. 48, fig 1, c), the liypapopliysis (ib., hy), and the neurapopliyses (ib. n, «), of the atlas. The centrum of the atlas is small, and has been anchylosed with that of the axis (PI. 48, fig. 2, x). Its vertical section is subquadrate, longitudinally grooved on each side; broader above than below. The hypapophysis (fig. 1, hy) is broader, but less deep, than the centrum, and the bases of the neurapophyses (ib., n,n,) extend down the sides of the centrum to articulate with the hypapophysis. The neurapophyses (ib., n,n) are ununited above, as they are below, and have yielded to the oblique pressure ; but the slight dislocation seems to have taken place without fracture of an upper union. There is no trace of spinous process ; but above the neural arch of the atlas is a pair of large, thick, transversely oblong, dermal bones or scutes ; the fi'actured surface of the most entire one, to the right (ib., r), is 2 inches in length by 1 in depth ; it exposes a compact peripheral texture of 4 or 5 lines in thickness at the upper and outer part, and about 1 line at the under or inner part, with a fine cancellous structure between. To the broad hypapo- physis of the atlas was articulated a long and slender rib {pleurapophijsis) (ib., In the foregoing constitution of the vertebral segment succeeding the skull we have the reptilian condition of the atlas, with modifications most closely repeated by the Crocodilia amongst the existing members of the class. The CrocodiUa alone show the transverse extension of the hypapophysis or "pseudo-centrum " of the seg'- ment, associated with the presence of articulations for the pleurapophysial elements. In lizards free pleurapophyses are not developed from the atlas or from the axis, rarely from the third cervical vertebra. But in Scelidosaurus the atlantal hypapo- physis is relatively broader than in CrocodiUa, and there is no trace of the detached representative of the neural spine which characterises the atlas in Crocouiha.^ In Plesiosaurus and Ichthyosaurus the true centrum of the atlas progressively acquires its form and proportions as such, and in the same degree resembles, in its relations to the basi-occipital, and to its own neural arch, the centrum of the first trunk-vertebra in fishes. The hypapophysis is proportionally reduced in size, and forms the first of the " sub-vertebral wedge-bones " in the Ichthyosaums? The second block of Lias (PL 48, fig. 2) includes the bodies of the axis {x) and of the third cervical vertebra, with parts of the pleurapophyses of the atlas {pi. a) and axis {pl.x) ; it includes, also, large, massive, dermal bones external and superior to the vertebral elements. The centrum of the axis is 1 inch 3 lines in length, from the line of adhesion of that of the atlas, part of which remains connected as the " odontoid process ;" the proper body of the axis is subcarinate 1 'Annals and Magazine of Natural History,' vol. xx, pp. 217 — 225. ^ Ichthi/opterygia, pi. xix, fig. 5, liy, a; pi. xsiii, fig. 2. 102 BRITISH FOSSIL REPTILES. below ; gently concave lengthwise at the sides ; compressed in the same degree at the middle, and slightly expanded at the extremities. The rib which it sup- ported {pV x) is shorter than that of the atlas, but, like it, is slender and straight ; about 3 inches of the atlantal rib is preserved, and about 2 inches of that of the axis. The body of the third vertebra presents a general increase of size ; it is 1 inch 8 lines in length, 2 inches 3 lines across the parapophyses at the fore part of the vertebra, 1 inch 6 lines across the posterior articular surface, and 1 inch 2 lines in depth. It is subcompressed at the sides, and more obtusely ridged below than the axis. The fore part of the body is articulated by an almost flattened surface with that at the back of the axis. The characters of the terminal articular surfaces were worked out more com- pletely in a consecutive cervical vertebra detached from the third block, and which, from its size, is probably the sixth. The part of the front surface (fig. 4, e) which is preserved is flat with a convex periphery ; the hind surface (fig. 3, c) is slightly concave, with a narrower and better defined circumference. The body of this vertebra is 1 inch 10 lines in length, 2 inches 3 lines across the parapophyses (p) ; 1 inch 8 lines across the hinder articular end (c). The under part of the centrum presents near its fore end a hypapophysial tuberosity ; it is constricted at the middle, and a small venous canal opens into that concavity on either side. The rib articulates by a bifurcate end with both par- and diapophyses ; the upper trans- verse process (fig. 4, rf) extends nearly 1^ inch from the neural arch; the neural canal («) is of a full oval form, with the small end downwards ; it is 9 inches in its longest diameter. The breadth of the neural arch, below the diapophyses, is 1 inch 7 lines. In the portion of the succeeding cervical vertebra, from the same block, the rib is directed more outwardly than in the antecedent one. The length of the neck of the rib is 1 inch 2 lines ; its thickness 6 lines, which increases after the development of the tubercle, where the fracture shows a subtriedral section. The poi'tion of the articular surface which is preserved of the centrum of the second vertebra indicates the same feeble concavity as in the preceding cervical vertebra (fig. 3). Supposing the vertebra (PI. 48. figs. 3 and 4) to be the sixth of the cervical series, it shows that the rib has more speedily resumed its normal character than in the Crocodilia. In these large existing Saurians the pleurapophysis, slender, straight and rather long, in the atlas and axis, becomes shortened and expanded in the third, fourth, fifth, and sixth cervical vertebrse, assuming in them a hatchet- like shape, with an overlapping arrangement; the posterior production of the "blade" lengthens in the seventh cervical; but the ordinary rib-shape is only resumed in the eighth vertebra, regarded as the first dorsal by Cuvier. LIASSIC DINOSAURS. 103 I infer, therefore, from the size and proportions of the two vertebras just described that they correspond with the sixth aiid seventh in the Crocodile, and that the Scelidosaurus, with probably other Dinosauria, differed from Crocodilia and from most Lacertilia in the long and slender form of most, if not all, of the cervical ribs ; but that these manifested their more essential Crocodilian affinity in their twofold articulation, by a bifurcate head, with distinct upper ( d ) and lower ip) transverse processes. The fourth block included, with the scapular arch, ten of the anterior dorsal vertebrae (Pis. 49, 50). The hinder fracture of the block has detached the anterior articular surface from the eleventh dorsal, the rest of which is the first of the series of the five following dorsals in the fifth block of Lias (PI. 52, figs. 1 and 2). The hinder fracture of this block has pretty equally bisected the last vertebra, which bears free ribs, viz. the sixteenth dorsal, the hinder half of which remains in the fore part of the block (PL 53), including the lumbar and sacral series of vertebrse. The section of the eleventh dorsal thus exposed near the anterior articular surface of the centrum is represented of the natural size in PI. 52, fig. 1, d ii. That through the middle of the sixteenth dorsal vertebra is similarly represented in fig. 2, D 16. The spinous process of the first dorsal vertebra (PL 45, D i) is 1^ inch in height and 8 lines in fore-and-aft extent ; the spine increases in both directions to the fifth of these vertebra3 ( 5 ), which is 2 inches 4 lines in height and 1 inch 10 lines in basal extent. The spines continue of about the same height to the tenth vertebra, D 10, with summits obtusely rounded, almost truncate. In the eleventh to the sixteenth dorsals, PL 51, d ii — d 16, the spines acquire their greatest fore-and-aft extent, with truncate summits, but no increase of height. Although these spines in the last six vertebras are nearly 2| inches in antero-posterior extent, their summits do not come into contact, but leave interspaces of from 5 lines to 8 lines. The prezygapophyses in the anterior dorsal vertebras look inward and a little upward, the postzygapophyses in the reverse directions, but as the vertebrjB recede in position the aspect of the surfaces becomes more nearly horizontal (PL 52, fig. 1 z). The diapophyses are subdepressed, 10 lines in breadth in the second vertebra, and gradually increasing to a terminal breadth of 15 lines in the ninth and tenth dorsals (PL 49, d, d). The parapophyses, as in the Crocodile, gradually pass from the centrum to the neural arch, and are seen at^, fig. 1, PL 51, upon the under and fore part of the diapophysis (d) in the eleventh of this series of dorsals, where the length of the diapophysis from the base of the neural spine is 2 inches 9 lines. No trace of parapophysis, or of the "head" of the rib, remains in the last three dorsals; the diapophysis is entire, as at d, fig. 2, PL 51. The ribs of the twelve anterior dorsal vertebra show both the head and the tubercle, the neck becoming gradually shorter in the last three. In the seventh 104 BRITISH FOSSIL REPTILES. vertebra the extent of tbe rib from the head to the tubercle is 2 inches 9 lines. In the tenth vertebra it is 1 inch 7 lines. The rib presents a shallow canal along its posterior surface ; it is nearly an inch in thickness. An extent of upwards of 10 inches of the body of the rib {pi, PI. 49) is preserved on the right side of this portion of the thorax of Scelidoiherium. The anterior dorsal vertebrae have been partially dislocated, especially the fourth from the fifth, apparently by pressure acting through the scapula (si) upon the diapo- physis and spine of the fifth dorsal. Beyond the scapula the vertebra have retained their natural position and connections, which seems to indicate the action of pressure whilst decomposition of the soft parts was going on in the carcass. Nine of the consecutive vertebra in the fourth block occupy the extent of 1 foot 9 inches. The breadth of the last of these vertebra (PI. 49, D lo), across the diapophyses ( ^), is 5 inches 4 lines. The total height of the eleventh dorsal vertebra (PI. 52, fig. 1) is 6 inches. The breadth of the centrum at the fractured part, near the anterior surface, is 1 inch 6 lines. The depth of the centrum, from the floor of the neural canal, is 2 inches. The breadth of the neural arch across what are called the " pedicles" is 1 inch 8 lines. The height of the neural spine is 2 inches 6 lines. As the vertebrae approach the sacrum the bodies gradually increase in depth, without g-aiuinff in breadth, until at the last dorsal the centrum, near its middle part, measures 2| inches in vertical and 1 inch 7 lines in transverse diameter ; a slight longitudinal impression on each side produces the contour of the transverse vertical section figured in PI. 52, fig. 2. The neural canal here gives a triangular section, with the apex downward and sinking into the substance of the centrum, but the sutural limit between centrum and neural arch are indiscernible. The diapophyses decrease in breadth and also in length, and now support the rib by a terminal, slightly notched, articular facet. The ribs, here with simple heads, become shorter and less curved ; a few, as in PI. 51, fig. 1, pi, have suffered fracture, with very little displacement. In different parts of the matrix of the blocks (Pis. 49 and 51) are portions of long and slender bones, which are, most probably, abdominal ribs. In the sixth block the hinder half of the last dorsal and one lumbar vertebrae are associated with the pelvis ; the lumbar vertebra (PL 53, l) had been dislocated downwards from its articulations with the sacrum. The four vertebrse of this part (PI. 63), with the iliac bones (62), are preserved almost in their natural relative positions, the sacral vertebrae having their neural spines and transverse processes exposed. Those of the first sacral (s, 1, rf, ;;/) stand out horizontally and transverse to the axis of the body ; a slight swelling {d), about one inch from their origin, may indicate the point of confluence of the pleurapo- physial {pi) with the par- and diapophysial elements of this part. It is A\ inches in length ; at its base it is 1 inch in thickness and 2 inches in depth, expanding in LIASSIC DINOSAURS. 105 that direction to fully 3 inches at the truncate extremity, and in breadth to 2 inches 2 Hues. Towards its end the process is excavated anteriorly, so that the rough terminal surface (fig. 2, pi.) abutting upon the iliac bone is reniform. Fracture of an angle at this surface in the left transverse process shows a medullary cavity of 10 Hues in diameter by 6 lines in the section as exposed, surrounded by a fine cancellous, almost compact, osseous texture, of from 2 lines to 4 Unes in thickness. The transverse processes of the other sacral vertebrjB gradually become shorter, with corresponding decrease of breadth at their origin, but with equal or greater expansion of their termination, that of the last {pi, « 4) measuring 2 inches 7 hues in fore-and-aft breadth ; the transverse processes thus touch each other, or nearly so, at their ends, and offer a continuous longitudinal surface for the ligamentous or fibro-cartilaginous attachments of the iliac bones (62) . The total length of the articular " sacro-iliac " tract, so formed, is about 10 inches; a very slight lateral twist or dislocation makes it rather longer on the left than on the right side ; this appears to have been due to great pressure after imbedding, and is accompanied by fracture or dislocation of the pleurapophysial part of the transverse process of the last two sacral vertebrge. The spinous process, in each of the four sacrals, is about 2 inches high and 2 inches 3 lines in fore-and-aft extent ; they touch each other by their rough, flattened summits ; these are narrow anteriorly, gradually expanding to a breadth of 8 lines at their posterior third, with a thick, rounded termination ; the position of these spines is over the interspaces of the origins of the transverse processes, through the backward inchnation or extension of the neural arches. The articular processes are faintly indicated at their base, the posterior processes overlapping the anterior ones of the succeeding vertebra. The longitudinal extent of the truncated summits of the four sacral spines is 9 inches. The hinder fractured surface of the block containing the sacrum exposes part of the first caudal vertebra, the rest being associated with the four consecutive caudals in the seventh block of Lias (PI. 54, figs. 1 and 2). The first caudal vertebra has been dislocated from the last sacral, and twisted half round, so that its spine lies upon the sacral transverse process ; the fracture has passed through the spine and part of the neural arch. The length of this spine (PI. 54, fig. 1, ns) from the upper part of the neural canal is 3 inches 6 lines, the transverse process (d) is 2 inches 3 lines in length, but its vertical thickness is reduced to 5 fines at 2 inches from its termination. The neural canal is 6 lines in breadth and 10 lines in depth. The bases of the neural arch seem to show that the anchylosis with the centrum had not here been complete. The length of the first caudal centrum (PL 54, c l) is 2 inches, the depth or vertical diameter of its articular end is 2 inches 5 lines ; the surface is moderately 14 106 BRITISH FOSSIL REPTILES. concave, with tlie circumference bevelled off convexly ; between tbe two expanded ends the centrum is moderately and uniformly concave lengthwise. There is no trace of hgemal arch in the first caudal. In the second that arch fPl. 54, fig. 2 h) is articulated to the posterior part of the under surface, and is produced into a spine of nearly 4 inches long. In the third caudal (ib., h, c 3) and succeeding ones the hgemal arch has been dislocated, showing its articular surface, which, by mutual union of the hgemapophysial bases, is single, sub-reniform, transversely extended, lightly concave across, and convex from above downward. The hasmal canal, thus circumscribed, and well shown in the fourth caudal vertebra, is about 2 lines in breadth and 1 inch 3 lines in length ; too narrow, it would seem, for the protection of the trunks of the blood-vessels supplying so long and so powerful an organ as the tail of the Scelidosaurus. This form of the hgemal canal or slit has every appearance of being natural, and not due to any posthumous compression. The hsemapophysial surface external to it is convex transversely, slightly concave lengthwise ; the lamina3 slightly contract to their union in the spine, which becomes compressed, and a little expanded from before backwards near its termination. The articular surface, after the second haemal arch, is afforded in equal proportions by the two conjoined centrums beneath their terminal junction. The transverse process of the second caudal (ib., d) arises fi'om the anterior two thirds of the vertebra, over the junction of the centrum with the neural arch ; a trace of the suture indicating the pleurapophysial character of this process is discernible in this and some following caudals. The length of the centrum is 2 inches 2 lines ; the fore-and-aft breadth of the base of the transverse process is 1 inch 5 lines ; its length is 2 inches 5 lines ; its terminal breadth is 10 lines, ending obtusely. The transverse processes progressively decrease in all these dimensions in the following vertebrte. The anterior zygapophyses (PI. 54, fig. 1, g) are twice the length of the posterior ones {z'}, by which their extremities are overlapped. The fore-and-aft breadth of the neurapophyses between these processes is 1 inch 2 lines ; that of the summit of the neural spine is 1 inch 6 lines ; the height of the spine from the base of the prezygapophysis is 3 inches 4 lines. These dimensions are taken from the third caudal vertebra. The five consecutive and coarticulated anterior caudal vertebrae in the present block of Lias give a collective longitudinal extent of 12 inches. The distal half of the right femur (PI. 54, fig. 2, 65), and parts of the right tibia (ib., 66) and fibula (ib., 67), are cemented to the vertebrae by the matrix. Figure 1 in this plate gives a side view, fig. 2 an oblique under view, of the first five caudal vertebrae. The succeeding (eighth) block includes the five vertebrae (PL 55, fig. 1) next in succession. In these the length of the centrum continues to be a little over 2 inches, but they gradually decrease in other diameters, and especially in the size of their diverging parts. The neural spine, in the ninth, is reduced to 2 inches 5 LTASSIC DINOSAURS. 107 lines in length ; the transverse process (ib., fig. 1, d) to 1 inch 3 lines. The haemal arch and spine retain a length of 3 inches 3 lines. That of the seventh vertebra (fig. 1, h) has a basal diameter of 1 inch 1 line, decreasing to 6 lines at the end of the neural canal, and thence to a terminal diameter of 2^ lines, the fore-and-aft diameter being here 10 lines. The centrums progressively become more concave and compressed between the articular ends. The prezygapophyses (ib., fig. 2, z) have their articular surface turned more inward, and grasp, as it were, the shortening rudiments of the postzygapophyses, the neural arch progressively contracting in breadth. The collective length of the five vertebrae in this block is 11 inches. The ninth block of Lias contains the five succeeding caudals (PI. 55, fig. 3). The centrums, exposed at their under and lateral parts, are singularly crushed, the sides of each having been pressed into the substance ; yet, where the cracks of the matrix expose the texture of the centrum, as in the fifteenth caudal (PI. 52, fig. 3), it shows a fine, compactly cancellous structure throughout; there is no trace of any such vacuity or unossified nucleus of the centrum as is met with in the vertebrge of Poildlopleuron, for example. The centrums retain their length of 2 inches. The hinder articular end of that of the tenth caudal (c) adheres to the fore part of the present block. In the next coarticulated vertebra, which is the eleventh of the caudal series (PI. 55, fig. 8, ll), the prezygapophysis (ib., fig. A, z) is 10 lines in length and 3 lines in breadth; the neural spine, measured from the base of the zygapophysis, is 2 inches in length ; the transverse process (fig. 3, d) is 1 inch in length, with half an inch of basal breadth. Nearly 2^ inches of the hsemal arch (ib., /;) are preserved. The pressure crushing the centrum of the eleventh vertebra has been applied to the middle of the under and lateral part; the articular ends have withstood, if they have received, it. The same is the case with the twelfth caudal. In the thirteenth the pressure has been more laterally appUed, and the outer wall, which has been driven in, preserves its vertical convexity. The diapophysis of this vertebra is 10 lines in length. In the fifteenth caudal (ib., 15, d) the diapophysis is reduced to 6 lines in length, with corresponding decrease of thickness. The five caudal vertebrge from the eleventh to the fifteenth inclusive occupy a longi- tudinal extent of 11 inches 6 lines. The tenth mass of Lias, fitting on to the foregoing, includes a consecutive series of nine vertebrge, viz. the sixteenth to the twenty-fourth caudal inclusive (PI. 56, fig. 1). In this series there has been a dislocation of the eighteenth from the nineteenth, and a similar one between the twenty-first and the twenty- second vertebrge, with an interval of nearly an inch between the separated articular ends of the centrums. These elements continue to decrease in vertical and transverse diameters, and also, but in a minor degree, in regard to their 108 BRITISH FOSSIL REPTILES. length. The transverse process has subsided to a tubercle upon the eighteenth (ib., fig. 1, d), and the postzygapophysis to a notch at the back part of the base of the neural spine, but the prezygapophysis (z) continues long and slender throughout this series. The neural spines progressively narrow and shorten, with a backward inclination. The base of the haemal spine (h) of the sixteenth caudal measures 9 lines ; its articular surface is transversely oblong. The surface for the articulation of the haemal arch, from this part of the tail onward, is chiefly afforded by the hinder and under part of its own vertebra, as in fig. 3, h. The hgemal arch and spine becomes reduced in the eighteenth caudal to the length of its centrum ; and in the twenty-third becomes shorter than the centrum, with a greater degree of antero-posterior expansion of the spine in proportion to the length of that part (ib., fig. 1,23,^). The transverse diameter of the anterior articular surface of the nineteenth caudal is 1 inch 6 Unes. The middle of the centrum has been reduced by pressure, attended vnth some fracture of the outer surface, to a diameter of 7 lines. In some of these vertebrae the middle, crushed parts of the centrum have been severed from the terminal articular expansions. I conclude, therefore, that they have been subjected to a general compressive force, probably connected with the change in the vertical relative position of the stratum. The compact layer of osseous tissue forming the articular end has resisted the pressure; the intervening, intermediate, cancellous structure has yielded to it. From three smaller portions of the matrix, succeeding the ninth block, eleven consecutive caudal vertebrae were wrought out, as in PI. 66, fig. 2, making us acquainted with a total of thirty-five caudal vertebrae of Scelidosaurus. In the last of this series the centrum (ib., 35) is reduced to the length of 1 inch, and the breadth of its fi-ont articular end to 6 lines. In the twenty- fifth caudal vertebra the ceuti'um (ib., fig. 3) is 1 inch 10^ lines in length, 1 inch 3 lines across the articular end, 7\ lines across the middle, the longitudinal concavity of the sides exceeding that of the under surface. At the fore part of this surface the hsemapophysial articulation is barely indicated ; at the back part it is marked by two surfaces {h), towards the most prominent part of which short, low ridges diverge. The low neural arch has coalesced with the upper three fourths of the centrum; the prezygapophyses {z) overhang the free fore part of the centrum, and extend beyond it to clasp the back part of the preposed neural spine. This is represented by a short, compressed ridge projecting above the part clasped by the prezygapophyses. The haemal arch of the twenty-fourth caudal (ib., fig. 2, 24, 7;) underlies the centrum of the twenty-fifth ; it presents a length of 1 inch 6 lines. Its closed base (ib., fig. 4) has a breadth of 7| lines; its presents a sub-bilobed form, concave transversely, convex from before backward. At the sides of the hsemal canal or rather slit, the arch has a fore-and-aft breadth of 4 lines ; the LIASSIC DINOSAURS. ]09 spine expands to twice tliat extent, witli an obtusely rounded termination. In the twenty-seventh caudal vertebra the haemal arch and spine are reduced to a length of 1 inch 2 lines ; the spine progressively decreases to the thirty-second vertebra, beyond which the hfemal element ceases to be developed. The centrum of the twenty-seventh caudal (PI. 56, fig. 5, 27) is 1 inch 10 lines long; the anterior surface is 1 inch in depth, 1 inch 2 lines in breadth. The coalesced base of the neural arch has an extent of 1 inch ; the prezygapophyses (j) are 9 lines in length ; the neural spine (««) is 1 inch in length above the zygapophysial surfaces, its summit penetrates the base of a superincumbent dermal bone, and the hfemal spine (h) has a similar relation to the dermal bone below. But both dermal bones may have been pressed nearer to the vertebra than in the living animal as the soft parts became dissolved away. The thirty- second caudal vertebra is 1 inch 4 lines in length, with a terminal breadth of 9 lines and a middle breadth of 6 lines. Its neural surface, showinsr the coalesced neural arch («), from which the processes have been broken away, is figured in PI. 56, fig. 6, the haemal surface is represented at fig. 7, with the last haemal arch (/;), which is not quite closed above. The thirty- fourth caudal vertebra (ib., fig. 8), is 1 inch 2 lines in length; the breadth of its front articular end is 7 lines. The anchylosed neural arch has a basal extent of 9 lines ; it is convex across the middle, like a saddle, rising into a short pyramidal process (ns) behind, like its peak ; and still giving off the pair of long and slender prezygapo- physes (?) from its fore part, which clasp the spine or peak of the antecedent vertebra. The thirty-five caudal vertebra, of which the principal distinctive characters have been above described, give a total length of 5 feet 8 inches 3 lines. The extent of dislocation between a few of these vertebrae would make a deduction of about 2 inches from the above extent ; but the few vertebra missing from the end of the tail, and reduced, as shown by parts preserved, to slender centrums, may, probably, have carried the length of the tail to about 6 feet. The trunk-vertebrge include, as has been shown, four sacral, one limabar, sixteen dorsal, and seven, or at least six, cervicals, and these vertebrae average each a length of 2 inches ; the total length of the vertebral column of the trunk, estimated as including twenty-eight vertebrae, would be, on the above average, 4 feet 8 inches, or, allowing for intervertebral soft parts, 5 feet at the utmost in the recent animal. The length of the head can scarcely have exceeded, more probably fell short of, 1 foot. Thus we obtain an appi'oximate estimate of the total length of the individual affording the before-detailed osteological characters of SceUdosaurus as not exceeding 12 feet from the snout to the end of the tail. But detached frag- 110 BRITISH FOSSIL REPTILES. ments of the fossilized skeleton of other individuals from the lower Lias of Charmouth indicate a larger size, and that the present is not that of a mature Scelidosaur. In the general osteological characters of the vertebral column we find this genus agreeing with Hijlceosaurus and Teleosaurus. None of the anterior vertebrae present the opisthocoelian modification charac- teristic of the Crocodilian genus Streptospondylus and in a minor degree of the Dinosaurian genera Chondrosteosaurus, Cetiosaurns, and Megalosmirus ; in this respect they come nearer to the vertebral type of the Iguanodon. Not any of the anterior dorsal vertebrae develops the spinous process of so disproportionate a length as they present in the carnivorous Megalosaurus. Although the neural arch becomes loftier than that of Crocodilia in the dorsal region of the spine, the exterior of the peduncles or neurapophysial laminae does not present the complex configuration produced by the strong, oblique ridges underpropping the diapophysis in Iguanodon and Chondrosteosaurus, Certain vertebrEe have a small unossified central part, relatively less than in Polypthy- chodon ; none show the extent of permanent chondrosal tissue characteristic of Chondrosteus. Upon the whole, I find the closest agreement to be between Scelido- and Hylreo-saurus in the characters of the vertebral column ; and I infer for both, but especially for Scelidosaurus, a greater aptitude for swimming than in the larger Binosauria. The scapular arch has been compressed transversely to a degree which has pi'oduced fracture of the right coracoid (PI. 49, 52'), without material displacement in its relations to the left (PI. 50, 52 and 52') » but with corresponding approxi- mation of the two scapulse (PI. 49, 50, 51 and 51')) which have squeezed together, with some fracture and more dislocation, the interposed parietes of the thorax. The right scapula (PL 49, 5l) is least displaced ; it extends along the first seven dorsal vertebrse, overlapping the spines of the last two. It is long and rather narrow; thickest above the humeral articulation, narrowest at its middle part, becoming broader and thicker towards its free end or dorsum, the margin of which describes a moderate and regular convex curve. The length of the bone to the fore part of its coracoid end is 13 inches; its least breadth is 2 inches; that of the base is 4 inches 10 lines. The body of the scapula describes a slight convexity outward in its course to the humeral joint, the expanded portion in front of which is gently excavated for a triangular space 4 inches long ; the apex being upward, with a well-defined boundary, indicative of the attachment of a muscle to this part. The anterior border is almost straight through three fourths of its extent from the base, then becomes slightly concave to the anteriorly produced angle of the coracoid end. The posterior border is more deeply concave, through the production of the thickened part of the bone to form the LIASSIC DINOSAURS. Ill humeral articulation (PI. 50, h). So mucli as is exposed of this surface is slightly concave transversely, slightly undulating in the opposite direction, 2 inches in breadth. The articulation (e) with the coracoid (PI. 49, 52') is a straight harmonia. At the upper part of the humeral articular process there is an oblong notch, with slightly raised borders. The left scapula (PI. 50, 5I) has yielded in two places to the external pressure, but without separation of the broken parts. It gives the same indication of the triangular muscular surface on the outside of its distal end as does the right scapula, the apex being defined by a better preseiwed, slightly raised, obtuse border. The fore part of the acromial end of the scapula (a), though fractured like that on the right side, is here better preserved, and gives a breadth of nearly 6 inches to this end of the bone. The humeral articulation (/,) measures 2 inches 6 lines, the coracoid one (c) 4 inches. A small, oval, dermal bone (rf), 1 inch 6 lines by 1 inch 3 lines, overlies the fore part of the scapulo-coracoid harmonia. It is flattened, slightly convex externally, like some others that seem to have defended the skin of the under surface of the trunk. The coracoid (ib., 52) is an almost circular, flattened, discoid bone, 5 inches in antero-posterior diameter and 4J inches in transverse diameter; the margin is most modified where it is expanded in two inches of its extent to contribute the coi'acoid portion ( ^ ) of the humeral joint. The scapular articular border ( « ) pre- sents less thickness. The mesial or sternal border ( m ), continuing the circular curve, touches its fellow ( 52) by only a small part of its circumference. The average thickness of the coracoid plate is 7 lines. About 1 inch 3 lines from the scapular surface there is a foramen, 5 lines in diameter. The free border of the entire coracoid appears to be raised, but this may be due to the included surface having been crushed in and cracked by external pressure. In the hinder intersjjace of the coracoids there is a flattened mass of a rhom- boidal form, composed of scattered portions of thin, dark, osseous substance, cemented together by matrix, which is discoloured by carbonaceous material. No part shows the continuous roughened, but compact, structure of the dermal bones. It appears rather to be the remains of some partially ossified element of the endo- skeleton. In its position it corresponds with the sternum. There is a fainter trace of the same kind of material, or discoloration of the matrix, at the anterior interspace of the coracoids. The humerus, which is preserved on the left side (PI. 50, 53 ), has been singularly crushed and flattened ; the side of the middle of the shaft being broken away, ex- poses a small medullary cavity. The distal end ( rf ) is broken off, and slightly over- lapped by the shaft (53). The length of the humerus is 11 inches 3 lines. It pre- sents a sigmoid flexure, the distal end slightly bent downward or forward ; the proximal articulationj moderately convex, is 3 inches 8 lines in the long diameter ; 112 BRITISH POSSIL REPTILES. the fore part is produced into a strong ridge, here partly broken away. The distal end is 5 inches across, and is moderately concave transversely behind. An osseous tubercle, 1 inch 4 four lines by 10 lines, is cemented to the anconal surface ; a second similar bone is attached to the interspace between the inner condyle and the slightly dislocated ulna. These are more probably parts of the scattered dermo- skeleton than tendinal sesamoids of the extensor of the forearm. The acromial end of the ulna (PI. 50, 55) presents a convex border 2 inches 2 lines in breadth. The mutilated head of the radius (ib., 54), preserving its natural relations to the outer condyle of the humerus, is 1 inch 6 lines in length. The shafts of the radius and ulna, with the rest of the bones of the fore limb, have been broken away. Four oval, dermal bones, like those overlying the humerus and ulna, are attached to the matrix in front of the humerus and radius. Behind the fractured sternal end of the right coracoid (PI. 49, 52) is the dislo- cated head (53) and anterior expanded pectoral process (p) of the right humerus (ib.), showing a thickness of 7 lines where it has been broken off. The transverse diameter of the humerus at this part is 6 inches, with a thickness of the shaft not exceeding 2 inches 9 lines, showing that the humerus in Scelklosaurus was more expanded and compressed proximally than in any existing reptile, and in this respect resembUng the same bone in the Dicynodonts.^ The proportions of the entire fore limb of Scelidosaurus, as indicated by the lencfth of the humerus, would be those of the same limb in Teleosaurus. The humerus is shorter than the scapula, barely equalling the extent of four coarticu- lated middle dorsal vertebrte. There is no trace of clavicle in the present speci- men ; the functions of the fore limb seem, therefore, to be less important in regard to locomotion on land than in Iguanodon, Megalosaurus, and modern Lizards. Yet the shape and proportions of the coracoid, as I pointed out in regard to the Stagonolepis when the remains and impressions of that reptile were submitted to my inspection by Sir Roderic I. Murchison, at Leeds, during the meeting of the British Association, September 24th, 18-58,' show the distinction from the Croco- dilian order and the afl&nity to the Thecodontion order and to modern Lacertilia, or give evidence of a more generalised reptilian character, in these extinct reptiles with dermal bones and scutes of the Lower Liassic and Upper Triassic deposits. 1 Op. cit., vol. i, p. 31, vol. ii, pis. 30—36. 2 Art. " PalEeontology," ' Encyclopaedia Britannica,' vol. xvii, p. 130, in which, in reference to the Elgin matrix of Stagonolepis, it waa stated that " no characteristic Devonian or Old Eed fossils of any class have been discovered associated with the foregoing evidences of reptiles, which, according to the determination of strata by characteristic fossils, would belong to the secondary or mesozoic period." LIASSIC DINOSAURS. 113 Pelvic Arch and Limb. Pis. 53, 57, 58. The left iliac bone (PL 53, figs. 1 and 2, 62) retains almost its natural relations with the sacrum. The right iliac bone (ib., 62) has been obliquely dislocated. It is a long bone, with a sigmoid flexure (ib., fig. 2, 62), convex upward and outward in its anterior two thirds, more slightly concave in the rest of its extent. Of the left ilium an extent of 18 inches is preserved, a part, apparently a small one, being wanting from both extremities. The narrowest portion of the bone is that which is produced anterior to the first sacral rib (ib.,sl); this portion is 6 inches in extent, triedral in form, 2 inches 6 lines in breadth where it joins the obtuse, expanded end of that rib. Beyond or behind the first sacral abutment the ilium progressively expands to a breadth of about 5 inches opposite the fourth abutment (« 4). The thickness of the bone, as exposed in the fracture of the left ilium, is from 2 inches to 2^ inches. The middle third of the substance of the bone shows a rather open, cancellous structure ; external to this the texture is much closer, with a compact, peripheral layer of from 1 to 2 lines in thickness. The articular cavity for the femur is on the under and outer side of that part of the ilium which is opposite its symphysis with the first two sacral vertebrse (s 1 and s 2). The fore part of the right ilium (g2') has been thrust away from that junction, and the femur (65) is dislocated, passing beneath the ilium, with the head abutting against the sacrum. The summit of the great trochanter terminates rather more than an inch below the articular head of the bone. The breadth of the femur across this part of the trochanter is 3 inches 6 lines. The length of the femur (PI. 57, fig. 1, 60) is 1 foot 4 inches. The inner process or ridge {t) begins to be developed about 5^ inches from the head of the bone, and is 2 inches in extent. The shaft of the bone at this part is rather flattened, both anteriorly and posteriorly, and is most convex externally. It assumes a rounder circumference about 1 inch below the inner process, where the bone is 2 inches 4 lines in diameter. Thence it expands to the condyles (a and j), becoming flattened anteriorly and concave posteriorly. The condyles are but feebly indicated by a shallow notch on the fore part, but more distinctly behind, where they are produced backward. The hind extremity of the outer condyle (j) is marked off by a notch from the rest of its articular surface lying anterior and external to it. This posteriorly defined part articulates with the outer condyloid production of the head of the tibia, the fibula articulating with the rest of the outer condyle. The transversely convex fore part of the shaft of the femur is divided on each side by a low ridge from the flattened surfaces converging towards it, the one from the outer side, the other from the inner process (t). The exterior of these ridges is continued further down the bone 15 114 BRITISH FOSSIL REPTILES. than the opposite one. This femur, being broken across about 6 inches from its upper end, shows a medullary cavity of about 1^ inch in diameter, with a compact and finally cancellous wall, which is nearly an inch in thickness next the base of the inner process, and is about 3 lines in thickness on the opposite side of the shaft (PI. 57, fig. 2). The transverse breadth of the shaft here is 2 inches 7 lines, the fore-and-aft breadth is 2 inches. The transverse breadth of the distal end is 4 inches 10 lines ; the fore-and-aft breadth of the outer condyle is 3 inches 3 lines, that of the inner condyle 3 inches 8 lines ; the depth of the posterior inter- condyloid notch is 1 inch 3 lines. The proximal ends of the tibia and fibula are crushed below their articular surfaces ; most so in the right leg, with fracture of both bones. The medullary cavity of the right tibia has been obliterated by this violence, and the strong, compact wall broken and crushed in upon it. The fibula, with a smaller cavity and thicker, compact walls, has better resisted the pressure, especially in the left limb. The length of the tibia (PI. 57, 66) is 12 inches 10 lines, that of the fibula (ib. 67) is about an inch shorter. The expanded upper end of the tibia passes over the outer and part of the front surface of the head of the fibula ;^ the expanded lower end of the tibia passes, in part, behind that of the fibula, showing a kind of twisted, terminally, overlapping relation between the two bones. There is a distinct interosseous space (o) between the upper three fourths of their shafts. The breadth of the proximal end of the tibia, which may be a little increased by compression, is 5 inches 6 lines. The breadth of the distal end is 4 inches 6 lines. The tibia, which, on the left side, has suffered least compression at its upper end, and has been partially dislocated from the femur, shows a coadapted surface of very similar shape to that of the femur, convex from before backward, slightly concave transversely at the back part of the joint. In both bones the articular surfaces are rough, as if they had been connected together ligamentously. The tibial articular sm'face divides posteriorly, as before noted, into two condyloid processes, with an inter-condyloid space of about 2 inches breadth : one condyle is for the inner condyle of the femur, the other is adapted to the posterior prominence of the outer femoral condyle. The back part of the proximal end of the fibiUa next the outer condyle of the tibia is similarly produced into a convex protuberance. The fore and outer part of the tibia is produced into a strong procnemial tuberosity or process. The shaft of the fibula contracts to a diameter of 1 inch 10 lines, and then expands transversely, but without corresponding fore-and-aft enlargement, to the distal breadth above recorded. 1 In their natural relative positions, the fibula has been slightly dislocated outward in the left leg. (PI. 57.) LIASSIC DINOSAURS. 115 To the major part of the distal end of the tibia, at least to two thirds of its inner or tibial side, is articulated the tarsal bone («), including the coalesced homologues of the astragalus, naviculare, with the ento- and meso-cuneiform bones, of the mammalian tarsus. This bone (PI. 57, fig. I, a) presents an anterior surface of an elongated, irregular, triangular form, with the apex tibiad or toward the inner side of the tarsus. It becomes narrower as it proceeds backward beneath the tibia (ib., fig. 3, a), its articular surface with which is concave from before backward, favom^ing flexion and extension, or motion of the foot to and fro. Its distal surface is convex in the same direction, and is sinuous transversely. The calcaneum (ib., figs. 1 and 3, Z) articulates with the distal end of the fibula (fig. 1, 67) and with the outer third of the same surface of the tibia (fig. B,l). The next most intelligible tarsal bone is that (figs. 1 and 3, i) which articulates with the calcaneum (f) and with the two outer metatarsals Uv and v). Its largest surface is tm-ned forward or upward ; its posterior surface is a smaller convex protuberance ; this bone answers to the cubo'ides. At the back part of the tarsus there projects the base of a wedge-shaped bone, (fig. 3, e) seemingly partially dislocated backward, which mainly supports the middle metatarsal (m), and extends partly over the fourth (fig. 3, iv). The apex of this bone appears on the front side of the tarsus (fig. 1, e) in the interspace between the astragalus {„), cubo'ides (j), the third and the fourth metatarsals. I regard this bone, therefore, as answering to the ecto-cuneiform ; I cannot discern any trace of other cuneiform bones, the fibro-cartilage by which the interspace between the bone («), and the first and second metatarsals, was most probably occupied held partly the place of the meso- and ento-cuneiform bones. From this it appeal's that the tarsus of Scelidosaurus includes but four bones, as in the Crocodile (PI. 58, fig. 4.) In the Lizard (Varanus), fig. 3, an ossification in the fibro-car- tilage at the base of the second metatarsal establishes the " meso-cuneiforme," and leaves the " ento-cuneiforme " to combine with the naviculare and astragalus in the bone (a). The metatarsus of Scelidosaurus consists of five bones. Of these, the fifth (PL 66, figs. 1 and 3, v) is abortive, and adherent to a rough ridge on the outer part of the base of the fourth metatarsal, with its proximal half extending over the interspace between tliat bone and the cuboid to articulate with the latter. It was not, however, flattened and expanded, as in the Crocodile (PL 58, fig. 4, «), but was slender and styliform, if we may judge by the proximal end which fortunately remains attached in the left hind foot of Scelidosaurus (PL 57, fig 1, »). It most probably did not support a toe, or make any distinct appearance in the entire foot. The other four metatarsals support each a fully-developed toe, with the progressive increase in the number of phalanges characteristic of sam-ian EejAiUa ; the first having 2, the second 3, the third 4, and the fourth 5 phalanges. 116 BRITISH FOSSIL REPTILES. The metatarsal of the first or innermost toe (Pis. 57 and 58, i), is 2 inches 9 lines long. With its proximal end laterally compressed, and abutting against the corresponding end of the second metatarsal, which is much expanded in that direction. The distal end of the first is 13 lines in breadth, with a convex articular surface. The first phalanx of the toe (») is 2 inches long, 14 lines across the base, convex transversely towards the dorsum of the foot, flattened transversely and slightly concave lengthwise, towards the sole. The ungual phalanx is 1^ inch in length, and 1 inch in basal breadth; is sub-depressed, and curved downward. About 3 lines in advance of the joint, its breadth is increased by two lateral ridges. The apex is subacute. The under surface is marked by many fine, wavy ridges. There are two obtuse longitudinal prominences on the under surface near the joint, for the advantageous insertion of the flexor tendons, and there is a rough prominence at the middle of the dorsal surface, near the joint, for the insertion of the extensor tendon. The dorsal surface near the margins and apex, is sculptured by vascular grooves. The total length of the first digit (i) is 6 inches. The second metatarsal (ib., a) is 5 inches in length; vdth a proximal articular surface 1 inch 6 lines in breadth, sinuous but almost flat : this surface presents almost double the transverse extent in the antero-posterior direction. The inner and anterior part of this surface is produced inward, or tibiad, apparently to afford an abutment or attachment, at least in part, to the proximal end of the first metatarsal. The outer or fibular side of the second metatarsal is almost straight, the inner or tibial one concave, the expansion at both ends taking place chiefly in that direction. The distal articular surface is convex from before backwards, with a median groove producing a transverse concavity between the two convexities or condyles, at the posterior half ; and these slightly project backward. The first phalanx of the second toe is 1 inch 3 lines in length, 1 inch 7 lines across the proximal, and 1 inch 6 lines across the distal end ; the diameter from before backward at the middle of the shaft is 6 lines, the phalanx is consequently broad and sub-depressed. The posterior or plantal surface at the proximal end is slightly produced. The distal articular convexity extends a little way upon the middle of the dorsal surface, and slightly swells out into two condyles at the opposite surface. The second phalanx is much shorter in proportion to its breadth, which at the base is 1 inch 6 lines ; the length being 1 inch 7 lines ; the tibial border is short and concave ; the fibular one is straighter and one third longer. The ungual phalanx {ii) differs chiefly from that of the first digit in its superior size, being 2 inches in length and 1 inch 4 lines in its greatest breadth ; the fibular margin is convex, the tibial one slightly concave. A side view of the bone, of the natural size, is given at fig. 4, PI. 57. The length of the third metatarsal (ib.. Hi) is 5 inches 4 hues. It is more sym- metrical in shape than the rest. The transverse breadth of the proximal end is LIASSIC DINOSAURS. 117 1 incli 8 lines ; the fore-and-aft breadth is 2 inches 1 line. The thickness in this direction diminishes rapidly towards the distal end ; the transverse dimension decreases in a much less degree ; this, at the middle of the bone, being 1 inch 2 lines, whence it increases to a distal transverse breadth of 1 inch 11 lines. The configm-ation of this articular surface resembles that of the second metatarsal ; the fore-and-aft breadth of the condyle is 1 inch 6 lines. The proximal phalanx of the third toe (ib.,m, l) is 1 inch 2 lines in length, 1 inch 10 lines across the base, and 1 inch 3 lines across from before backwards. On the middle of the outer border is a tuberosity ; each side of the distal end is deeply impressed ; the distal articulation resembles that in the second toe. The greatest transverse breadth of this phalanx is 1 inch 7 lines. The second phalanx (ib., 2), with a basal breadth of 1 inch 6 lines, is only 1 inch 7 lines in length. The distal articulation is 1 inch 5 lines in breadth. The third phalanx (ib., .3), with a basal breadth of 1 inch 3 lines, is 1 inch 2 lines in length, with a distal breadth 1 inch 1 line. The ungual phalanx (ib. 4) is more depressed in proportion to its breadth than that of the preceding toe ; in other respects it resembles it in shape. The fourth metatarsal (ib., iv), is 4 inches 5 lines in length, of an unsymmetrical figure, receding from the middle metatarsal along its distal half, which is concave lengthwise on the tibial side ; the fibular side presents a general but slighter con- cavity; this metatarsal is triedral, the fore and back surfaces converging to an obtuse, narrow, outer border, significant of its terminating that side of the foot beyond the representative style of the fifth digit (y). The fourth metatarsal mea- sures 1 inch 9 lines across the base and 1 inch 7 lines from before backwards, at the tibial side of the base ; the fibular side being reduced to the narrow rough ridge for the ligamentous attachment of the fifth abortive metatarsal. The breadth of the shaft of the fourth metatarsal at its lower third is 1 inch 1 line ; that of the distal articular surface is 1 inch 5 lines. The first phalanx of the fourth toe is 1 inch 11 lines in length ; the basal breadth is 1 inch 8 lines and the distal breadth is 1 inch 6 lines. The tibial angle of the proximal surface is most produced. The fore-and-aft dimensions of the shaft do not exceed G lines. The second phalanx is 1 inch 2 lines in length, and 1 inch 3 lines in basal breadth. The third phalanx is ] inch in length, 1 iucli 4 lines in breadth ; the fourth phalanx is 9 lines in length, 1 inch 2 lines in breadth. The ungual phalanx is 1 inch G lines in length ; 8 Hues across its articular surface, 11 lines across its broadest part, caused by the aliform expansions of the bone beyond the articulation. It curves downwards and inwards, or towards the tibial side, to a subacute apex ; the characters of its sui'face corre- spond with those of the larger ungual phalanges of the preceding toes. From the abortion of the fifth digit, and the disproportionate shortness of the first, we have in Scelidosaurus the example of a reptile manifesting a tendency to the tridactyle type of the hind foot, and this is eflfected in its remote successor of 118 BRITISH FOSSIL REPTILES. the Wealden period — the Iguanodon — by the suppression of the first, and by a similar atrophy of the fifth digit. The foot-prints of Scelidosaurus would termi- nate forward by the marks of four claws, the innermost falling short of the base of the second, this and the fourth reaching the same line, and the intermediate third claw extending farthest. The hind foot-prints of Iguanodon are tri- dactyle.^ The total length of the foot of Scelidosaurus is 1 foot 1 inch 6 lines ; the length of the leg (" cnemion ") is 1 foot ; the length of the thigh is 1 foot 4 inches ; consequently the total length of the hind limb is 3 feet 5 inches ; and, allowing for the fibro-cartilaginous matter of the joints and the terminal claws, the limb may have been 3 feet 8 inches long in the recent animal. The femur equals the length of about seven co-articulated dorsal vertebrce, and, with the leg, manifests longer proportions to the body than in the CrocodiUa; but the foot presents shorter and broader proportions, although it has the same number of toes. Scelidosaurus, however, difi'ers from Teleosaurus and modern CrocodiUa, in retaining the ungual phalanx of the fourth toe, as in modern hzards (PI. 58, fig. 3, to) ; although it differs from these and resembles the Crocodiles in the non-development of the fifth toe. The interesting evidence of this intermediate relationship aiforded by the bones of the hind foot, as by some other parts of the skeleton, is illustrated by the outline figures of the skeleton of the hind foot (PI. 58) in Varanus, fig. 3, in Grocodilus, fig. 4, and as similarly restored in Scelido- saurus, fig. 2. In the same plate is figured, of half the natural size, the bones of the right hind foot of the skeleton of the Scelidosaur which has yielded the subjects of the present Monograph ; showing the effects of pressure in fracturing and partially dislocating the metatarsal segment, after all the joints of the toes had been cemented by the surrounding hardened matrix in their respective varied numbers and co-adjustment in each toe. Bermo-slceleton. The bones belonging to this system were extensively developed in Scelido- saurus, and are for the most part of a massive character. They have been much displaced in the present specimen, partly during the decomposition of the carcass, and partly by subsequent pressure due to movements of the imbedding stratum ; but retain their most intelligible natural relations to the endo-skeleton in the caudal region : in which part, therefore, I shall begin their description, as they were found, on exposing the vertebral characters on the left side, from the end of the tail forwards ; and were either removed, or left in situ, as the case required. ' Atife, Vol. i, p. 374, Pis. 43, 44 {Dinosauria). LIASSIC DINOSAURS. 119 At the thirty-first caudal vertebra, for example, there was attached to the back part of the neui^al arch, and pressed rather obliquely to the left side, an elongated triedral dermal bone, with the narrowest side or surface forming the base, and the two broader or larger lateral surfaces converging at an acute angle to an upper ridge. Much of this ridge on the fore part of the bone had been broken away in the original exposure of the specimen ; the length of what remained was 1 inch 2 lines, with a basal breadth of 6 hues. The sides of the bone seemed as if worm-eaten, by narrow curved grooves with intervening small, oblong, and cir- cular pittings. The texture as exposed by the fracture was compact, reflecting a lustre. Between the twenty-ninth and thirtieth vertebrae there was the basal part of a similarly shaped dermal bone, 1 inch 9 lines in extent, with a basal breadth of 9 lines. It lies upon the right side of the co-adapted halves of the neural arches of these vertebrae, but may have been displaced from the median line, and this is more probable as the base of a dermal bone crossing the articulation between the centrums of the same caudal vertebrte has also been pressed towards the right side, on which the carcass of the reptile appears to have rested in the matrix. But any doubt as to the relations of the dermal bone above indicated was dissipated by the better preservation of those found in connection with the twenty- seventh and twenty-eighth caudal vertebras (PI. 56, fig. 2), and which are represented of the natural size in figm-e 5 of the same plate. The dermo-neural bone (dn), was found fractured, with a slight displacement of the back part of its base : when entire, it had a longitudinal extent of 3 inches 6 lines, and a vertical one of 2 inches. The base is hollow, and has been crushed by the lateral pressiu-e ; but seems to have had a breadth of nearly an inch. The sides converge to the upper margin, which describes a bold convex curve from before backwards, along two thirds of the contour, and then descends in a straighter hne obliquely backward to the hinder angle of the base. This dermal bone extends from above the prezygapophyses {z) of the twenty-seventh caudal vertebra to the fore part of the spine of the twenty-eighth. On removing part of the side of the base of the dermo-neural bone the spine (ns) of the twenty- seventh vertebra was seen to have penetrated the basal cavity, as far as that extended into the substance of the dermal bone ; but I incline to think that fibrinous or other soluble tissue intervened in the living reptile, and that the position of some of the more anterior dermo-neurals, situated at a higher level above the neural spines, was the more natural one. The dermo-h^mal bone (ib., d h) presents a longitudinal extent of 2 inches 3 lines, with a vertical one of 13 lines, and a basal breadth of about 9 lines. The haemal spine of the twenty-seventh vertebra (h) seems also to have entered a hollow in its base, where it was exposed by removal of part of the left wall of the basal cavity. But this had been pressed up to the under part of the 120 BRITISH FOSSIL REPTILES. centrums, almost touching the posterior half of the twenty- seventh and the contiguous two thirds of the twenty-eighth caudal vertebra ; obliterating an interspace which should have been occupied by muscle, tendon, ligament, and other soft parts in the recent animal. The dermo-h^emal spine below the twenty-fifth and twenty-sixth caudals differed only in its larger size from the succeeding one. Part of the base of the corresponding dermo-neural was preserved. In the series of nine consecutive caudal vertebra (PI. 56, fig. 1), the number and disposition of the dermo-neural and dermo-h^mal bones were more fully and satisfactorily exhibited. Three consecutive dermo-neurals extended over a series of seven vertebrae, from near the fore part of the first to near the hinder half of the last of these seven, each extending over the interspaces of two vertebrce. The corresponding dermo-h^mals are of smaller size, cross only one intervertebral space, which is the second or posterior of those so crossed above, and their hinder end is a little further back than that end of their homotype above. But, in working out these vertebriB, indications of a third series of caudal dermal bones were first met with. There extended over the articulation between the twenty- first and twenty-second caudals the base of a dermal bone, 3 inches long, crushed, with its apical ridge broken off. On its removal, the vertebrae it crossed were seen to have been displaced to the extent of nearly an inch. The position of this bone, and the ascertained relations of the neural and hfemal dermal bones to their vertebrae, made it improbable that it was one of either of these series displaced ; and attention was quickened, which led to the detection of a similar appearance further in advance, to be presently described. The best preserved dermo-neural, in the series of nine caudal vertebrae (PL 56, fig. 1, dn), presents a basal longitudinal extent of 3 inches 5 lines, with a basal breadth of 1 inch 9 lines ; its quasi worm-eaten, rugose sides, converge to an upper margin, not quite entire, but with apparently a contour resembling the dermo-neural in fig. 5. The present larger bone overlies the twentieth and contiguous portions of the nineteenth and twenty-first caudal vertebrae. The corresponding dermo-haemal bone {d, Ji), with a longitudinal basal extent of 2 inches 6 lines, and a basal breadth of 1 inch 3 lines, underlies the twentieth and twenty-first caudals, extending along a greater proportion of the former. Its sides, similarly but more finely sculjjtured than the dermo-neural above, converge to a convex inferior border ; the depth of the side being not less than 1 inch 6 lines. The next dermo-neural in advance overlies the eighteenth and contiguous half of the seventeenth caudal vertebrae. It presents a basal extent of 3 inches 6 lines, with a basal breadth of 1 inch 6 lines. The base of the corresponding dermo-htemal spine is preserved, which underlaps the hinder two thirds of the eighteenth and the front third of the nineteenth caudal. Its base is LIASSIC DINOSAURS. Ui 2 inclies 7 lines in length, with a moderate contour. The apical ridge and left side of this bone have been broken away. Between the above-described dermo-neural and dermo-h^mal bones there was the base of a lateral dermal bone, 3 inches 5 hnes in length, applied over the eighteenth and part of the nineteenth caudal vertebrae, like that between the twenty-first and twenty-second. The portion preserved in exposing these vertebras is figui'ed in the interspace produced by their slight dislocation, into which it had been wedged by pressm-e. I conceive it to have been the direct instrument of the dislocation, receiving and transmitting the extraneous pressure ; and at a period when the vertebrse in front and behind were sufficiently free in their bed to allow of being pressed close together, with obliteration of their .natural interspaces originally occupied by the soft inter-articular material ; the extent of such inter- space is probably shown between the twenty-second and twenty-third caudals (PI. 56, fig. 1). From the evidence of the dermo-neurals and dermo-htemals, in sitif, in the present series of vertebrae, the dermal bone above described could not be one of these series displaced ; and I infer from it, and the evidence of a similarly situ- ated bone in a remoter part of the tail, that this appendage was defended by a series of lateral as well of upper and lower dermal ossicles, though, perhaps, in less number, and of a flatter figure, along the sides. The next dermo-neural in advance overlaps the sixteenth and the contiguous half of the fifteenth caudal vertebrge ; but its hinder end, as well as a part of its summit, are broken away. What remains, measures 3 inches 4 lines in length with a basal breadth of at least two inches. The margin of the base of all the above- described dermo-neurals describes a gentle convexity. As the dermo-neurals advance in position, they progressively acquire increase of basal breadth, to near the base of the tail, retaining the avei'age length of 3^ inches, with a small increase of height. Three dermo-neurals range along an extent of the five vertebrae (eleventh to fifteenth caudals) figured in PI. 55, fig. 3 ; and the same relative number and position are shown in the five antecedent caudals (ib., fig. 1, dn). On the right or imbedded side of the vertebrae, overlying the centrum of the fourteenth, and contiguous parts of the thirteenth and fifteenth vertebras, is the base of a dermo-lateral bone, 3 inches 3 lines in length, 2 inches 2 lines in breadth, the sides converging at an open angle, but with their terminal ridge broken off. This representative of the lateral series of dermal bones would seem to show that they had greater breadth and thickness than either those of the upper (neural) or lower (hifimal) dermal series. The right side, where these additional indications of a lateral series of dermal bones are preserved, was that which was left imbedded in the matrix ; the left side being that which was exposed by the original quarry- ing operations. It is probable, therefore, that the dermo-lateral bones of the left 16 122 BRITISH FOSSIL REPTILES. side, with tlie exception of the few remains above noticed, were in the matrix so detached. The characters of the caudal vertebrsB figured in Pis. 55 and 56 were displayed by careful removal of the matrix left adhering to the parts originally exposed ; during which operation the portions of the dermo-lateral bones which had been pressed inward, and contributed to the dislocation of the twenty-first from the twenty-second, and of the eighteenth from the nineteenth caudal vertebras, were brought to lig'ht. A dermo-neural bone overlies the ninth and tenth caudals (PI. 55, fig. 1) ; another over the seventh and eighth [d n) ; a third over the sixth and fifth. The fracture through the middle of this latter bone (PL 54, fig. 3), shows the form and depth of the angular excavation at its base, which rested, probably with interposed ligamentous substance, upon the summit of the neural spine of the caudal. The corresponding dermo-heemal bones, displaced so as obliquely to overlap the h^mal spines on the right side, are also preserved ; and on this side there are as many dermo-lateral scutes, but more fragmentary and dislocated. In the block of lias with the first caudal vertebrse (PI. 54, figs. 1 and 2), is the anterior half of the dermo-neural overlapping the fifth and sixth of that series. Two similar bones with a basal excavation exposed by fracture in one of them, are situated to the right side of the fourth and third caudals, which may be dermo- laterals or displaced dermo-neurals. A portion of a massive dermal bone lies upon a part of the ilium contained in this slab. The rest of the armour of this part of the base of the tail has been removed. The like is the case with regard to the upper part of the block including the sacrum (PL 55). At its under part, in which are imbedded dislocated bones of the hind limbs, there are a few scattered portions of wedge-shaped dermal bones, similar in size to those at the base of the tail, but less pyramidal, and with more obtuse summits. A few smaller, flatter, subcircular dermal bones were met with in the course of exposing the parts of the endo-skeleton. One of these (ib., d), lies above the interspace between the left ilium and the third sacral rib (PL 55, fig. 1, d). In the block of lias containing the fore part of the thorax and scapular arch a longitudinal series of eight dermal bones were found on the right side, overlapping the ribs, external to the diapophyses. These dermal bones were shorter and thicker than the caudal dermo-neurals, and had been subject to more or less fracture and some displacement. The best preserved was wedge-shaped, with the sides of the excavated base slightly convex, 2 inches in length, 8 inches 9 lines in breadth, the sides converging at a more open angle, but unequally, to a margin which shows a convex ridge. The inferior size and unsymmetrical shape of this bone seem to show that it formed part of a lateral row, which had been situated near a middle one, or had ranged along near the medial line of the back. The margins of these bones wei-e not entire. The summit of a dermo-neural spine remains wedged LIASSIC DINOSAURS. 123 between the spines of tte second and third dorsals, and another between those of the fourth and fifth dorsals (PI. 49, dn, dn). On the left side of the thorax (PL 50) are preserved some of the upper lateral series of dermal bones {dnl), showing their natural position and intervals. On the same side, beneath the fore- going (PL 50, d l), are some larger wedge-shaped dermal bones. Three of these may have been displaced from above the neural spines. They are elliptical, 3 inches long, 2 inches broad at the base, with the sides converging with a slight concavity to the upper ridge, which has been broken off in each, so that its height is conjectural. Other evidences of dermal bones on the under part of this slab are too fragmentary and scattered to throw any light upon their natural arrangement. On the right side (PL 49), overlying the ends of the ribs, about ten inches distant from the vertebrae, are preserved three of a series of flattened, sub-ovate, dermal scutes ( da, da), about 3 inches by 2 inches in the long and cross diameters, and from 2 to 4 lines in thickness. The outer surface exhibits the same character of sculpturing as do the dermal bones of the tail ; the inner surface is smooth. In the block containing the second and third cervical vertebrne the pair of lateral, unsymmetrical, dermal bones have been preserved nearly in their natural position. They are three-sided ; the shortest is dii-ected mesiad ; the side next in length looks downward ; the outer surface, more convex, is directed upward and outward, and is the most extensive. These scutes have been fractured through their centre. They show an external, very compact, layer of bone thickest on the outer or peripheral side. The rest of the bone shows a rather close cancellous structure. Above these, but slightly displaced, is a pair of wedge-shaped bones, which are probably dermo-neurals, indicative of a parial arrangement of these along the nape, contrasting with their single series above the tail. Each of these dermal bones are somewhat unsymmetrical in form, 2 inches 9 lines in the length of the base, 1 inch 9 lines in breadth, with the median surface more extensive than the outer, and both converging to a ridged summit, but which is broken away. The anterior pair of nuchal scutes is preserved in connection with the occiput, overlapping the atlas (PL 48, fig. 1, dn, r ). They are similar in shape, but smaller in dimensions, than those last described, and have been broken across. From the sum of the foregoing observations, it may be inferred that the surface of the Scelidosaur was defended by several longitudinal series of massive dermal bones, those occupying the median and upper surface being arranged in pairs upon the nape and singly along the tail. External to these were a lateral series at least two in number, but probably more, on each side the trunk, having the same wedged and ridged shape as the dermo-neiu-als. Beneath these were flattened, ovate scutes along the lower lateral part of the thoracic-abdominal region. In the tail we have more decisive evidence of a single median row of large, symmetrical, 124 BRITISH FOSSIL REPTILES. cuneifoi'm, hollow-based, superiorly ridged dermo-neurals, with dimensions making three occupy the space of five vertebrEe along the base of the tail, and nearly seven vertebra} along the hinder half of the tail. There was a corresponding median series of smaller and less vertically extended dermo-haemal bones, and also a single series of dermo-laterals, of more depressed and fuller ovate form, on each side. The accidents attending the decomposition of the carcass of this reptile seem to have had the chief share in the removal and displacement of so large a proportion of its coat of mail. Subsequent cosmical violence has been concerned in the fracture, the crushing, and in a certain amount of displacement of the constituent parts of the skeleton. Lastly, further fracture of the fossil bones has been due to the quarrying operations, by which the specimen was brought to light. A few remains, including a femur, 5 inches in length, but with both extremities seemingly incompletely ossified, indicated a young Dinosaur ; and with characters, as of the inner trochanter, in regard to shaj^e and relative position, which led me to surmise that it might be part of an immature and very young individual of a Scelidosaurus. These remains were from the same liassic locality as the larger, probably adult bones. To whatever extent the Saurian organization has been modified for terrestrial life, that has been, in no instance, such as to suggest an inability to swim. On the contrary, the disproportionate shortness of the fore limbs, even in Iguanodon, leads to the suspicion that they might be short in reference to diminishing the obstacles to propelling the body through water by actions of the strong and vertically extended tail; and that, as in the living land lizard (AmUyrhynchus), of the Gallopagos Islands, the fore limbs might be applied close to the trunk in the Iguanodon, when it occasionally sought the water of the neighbouring estuary or sea. One would suppose that the newly born or newly hatched young of a Dinosaur might be safer on shore than at sea, or at least in waters which, like those of the Liassic ocean, seem to have swarmed with carnivorous Bnaliosaurus. If the Dinosauria were ovo-viviparous, and jDroduced but few young at a birth, the remains from the lower Lias, noticed at p. 90, might be those of a foetus borne by a gravid Scelidosaur to sea during an occasional excursion, and which by some casualty had there perished, and become imbedded, with her progeny, in the muddy bottom of the old Liassic ocean. I have not, however, been able to obtain precise evidence of the proximity of the small femur with the larger one of the Scelidosaurus, and bones of more than one small individual might have been expected to occur in juxtaposition if they had perished before birth. The analogy of the crocodile, moreover, would lead us to expect that the newly excluded or newly born Scelido- saur would be of smaller size than the individual indicated by the bones first discovered. LIASSIC DINOSAURS. 125 The general condition of the almost entire skeleton of a Scelidosaur organized, as seems by the structm-e and proportions of the hind foot, for terrestrial rather than aquatic hfe, or at least for amphibious habits on the margins of a river rather than for pursuit of food in the open sea, led me to infer that the carcass of the dead animal had been drifted down a river, disemboguing in the Liassic ocean, on the muddy bottom of which it would settle down when the skin had been so far decomposed as to permit the escape of the gases engendered by putrefaction. In that predicament the carcass would attract large carnivorous marine fishes and reptiles, and portions of the skin, with prominent parts not too strongly attached to the trunk, would probably be torn away before the weight of the bones had completely buried the carcass in the mud. In this way, perhaps, we may account for the loss of much of the dermo-skeleton and of the two fore feet. The larger hind limbs with their stronger muscles and ligaments, would offer better resistance to such predatory attacks ; and they, at any rate, have been preserved. The agitation to which the body must have been subject in its course down the stream, and before it finally sunk and settled out of sight, would be attended, after a certain amount of decomposition of the flesh, ligaments, and other soft parts, with such an amount of dislocation as the ribs and other parts of the vertebral column exhibit along the otherwise well-preserved and completely consecutive series of the bony segments, from the skull to near the end of the tail. But the oblique compression of the skull, the flattening of the thorax, squeezed between the approxi- mated piers of the scapular arch, attended with fracture of one of the coracoids, and other indications in the rest of the trunk, plainly bespeak the enormous pressure to which the fossil has been subject after its imbedding, and which must have been attended with stiU more injury and destructive obliteration of anatomical characters had it not been for the surrounding uniform support afforded by the matrix, compactly hardened around the petrified skeleton before those cosmical movements commenced to which the change in the position of the old Liassic sea-bottom has been due. 126 BRITISH POSSIL REPTILES. SUPPLEMENT No. II. MESOZOIC LIZARDS. Genus — Echinodon,'' Oiven. EcHiNODON Becclesii, Owen. ' Lacertilia,' PI. 11, figs. 1 — 9. The specimens figured in the above-cited plate were discovered by S. H. Beccles, Esq., F.R.S., in the thin, fresh-water stratum at Durdleston Bay, Isle of Purbeck. They consist of portions of the upper and lower jaws of a Saurian, allied, by the shape of the teeth, to Macellodon (PI. 11, fig. 10, a—e),^ but of larger size, and with the thecodont implantation of the teeth. The crown belongs, in general shape, to that type, of which the teeth of Palceosmirus, Scelidosaurus, Gardiodon, HylcBOsaurus, and even those of Iguanodon, are modifications. The teeth of the present genus are distinguished by the marginal serrations of the apical half of the crown, which increase in size from the apex to the base of that angular part of the tooth, the two basal points resembhng spines, and terminating respectively, or forming the confluence of, the two thickened ridges (ib., r, fig. 2, c) bounding the fore and hind borders of the basal half of the crown. The crown is supported on a subcylindrical fang, and suddenly expands, both transversely (PI. 2, fig. 11, c) and antero-posteriorly (ib., j). In the former direction it as quickly begins to contract, and the outer and inner sides converge in almost a straight line to the apex ; in the latter direction the crown continues expanding for about half, or rather more, of its longitudinal extent, with a shghtly convex contour; it then rapidly contracts to the apex, the converging borders meeting at a right or somewhat acute angle, and being serrated as above described. The thickest mid-part of the crown forms a longitudinal rising, usually more marked on one side of the tooth ; at the apical half the crown gradually becomes thinner towards the fore and hind margins ; but at the basal half these margins are thickened, and cause the surface between them and the mid-rising to be undulated transversely. At the apical part of the tooth both the outer and inner sides are gently convex, the transverse section giving the thin-pointed ellipse, as in fig. 6, b. The outer and inner enamelled sides of the crown each describe a curve at their base (fig. 3, b,r), convex towards the fang; these bases are somewhat thickened 1 'Ex'tt'os, hedgehog, and oSows, tooth, "prickly tootb." 2 ' Quarterly Jouraal of the Geological Society,' No. 40 (1S54), p. 422. MESOZOIC LACERTIANS. 127 and rounded, so as to project from the fang ; they converge at the fore and hind parts of the tooth, and unite at an acute angle (fig. 2, c, r), to form the long, basal points (fig. 3, i, s) of the serrated half of the crown. The foregoing characters apply to the majority of the teeth of EcMnodon, A portion of the left maxillary bone, with its outer surface exposed, is repre- sented in PL 11, fig. 1, and in outline, of the natural size, at «. The anterior, probably premaxillary, part has been detached and broken. Three teeth, more or less fractured, project from sockets in the alveolar border of this part; their crowns are less expanded than in the typical maxillary and mandibular teeth. Part of the boundary of an external nostril is indicated at «, the larger maxillary fragment the first two teeth present a similar form, and the entire crown of the second shows it to be longer, as well as more slender, than the posterior teeth ; it resembles a canine tooth in both shape and position, the crown being subcom- pressed and slightly recurved, as well as sharp-pointed. It would serve well to pierce and retain a living prey. It recalls a dental character of Iguanodon. The tooth succeeding the laniariform one presents the typical characters ; In fig. 1 are shown the impressions of four of the teeth preserved in the slab (fig. 2). Above the first impression (o, fig. 1) is the crown of a successional tooth, about to displace the tooth (o, in fig. 2). The outer side of maxillary teeth is shown, magnified, at i,b. The remainder of the upper maxillary, with part of the palatine and pterygoid bones of the left side, are represented, magnified, in fig. 2, and of the natural size, in outline, at „. The extent of the inner alveolar wall, effecting, with the cross partitions, the lodgment of the teeth in sockets, is here demonstrated. The expanded crowns of the teeth come into contact. The inner surface of the crown is shown at b, in which the middle longitudinal rising is rather less prominent than on the opposite surface. The fore part of the crown is represented at e. The outer side of a portion of the right maxillary, with eight contiguous molars, is represented in fig. 3, and of the natural size, in outline, at a. There is a Unear row of small foramina above the alveolar border. The median longitudinal rising of the crown of the teeth is more strongly marked on this, the outer surface. In fig. 4 is represented the inner surface of the posterior part of a right maxillary, containing six contiguous teeth, with a less prominent or less defined median rising of the teeth in this fragment ; the last three teeth gradually decrease in size. The inner surface of a portion of a mandibular ramus, with eight contiguous teeth, is represented at fig. 5, and in outline, of the natural size, at a. The fore part of a right ramus, consisting chiefly of the dentary element, is represented in figs. 6 — 8, and of the natural size, in outline, at «. Fig. 6 gives the outer side, but the whole vertical extent of the bone is only preserved at the symphysial end. The apex of a young tooth projects from the fifth of the sockets here preserved ; 128 BRITISH FOSSIL REPTILES. it is represented magnified at a and b. There is a linear series of small nervo- vascular foramina a little below the alveolar border. The crowns of the developed teeth have been broken away ; their fangs in the sockets are shown in fig. 7 ; the anterior teeth are narrower than the rest, as in the upper jaw. On the inner side of the specimen (fig. 8), a considerable extent of the symphysis {s, s) is shown. The posterior part of a broken dentary element of the left ramus is represented in fig. 9, showing the last eight teeth, and the impressions of the crowns of as many in advance. A portion of the crown, displaced, of the fourth from the last is preserved, and likewise portions also of those in advance, which have been broken in splitting the slab, so that they appear smaller than they actually were. The last three teeth are entire, and show a gradual decrease of size, as in the portion of upper jaw (fig. 4). A magnified view of the inner surface of the last lower tooth is given at a, fig. 9. The reference of Echinodon to the Lacertians is suggested by its diminutive size and by certain characters of jaws and teeth, but the structure of the vertebrae and limb-bones must be ascertained before the ordinal afiinities of Echinodon can be satisfactorily determined. The modifications of the mode of implantation of the teeth in the known limits of the Dinosaurian order affect the value of the thecodont character as a mark of affinity. THE FOSSIL REPTILIA OF THE LIASSIC FORMATIONS CHAPTER TV. Order— CROCODILIA. Of the fossil evidences of this order of Reptiles, represented by the existing genera Grocodilus, Alligator, and Gavialis, those referable to such genera have not hitherto come to my knowledge from strata older than the tertiary formations (vol. i, pp. 80—129). Vertebrae and other remains from Cretaceous series, as the Green-sand of Sussex, show vertebral modifications of higher than generic value, and upon these, with dental and cranial characters, have been founded the extinct genus Gonio])holis {ante vol. i, p. 199). Similar parts of the osseous and dental system, from "Wealden formations, have yielded characters of the genera Streptosijondylus (p. 398), Suchosaurus (p. 433), and Eylceochampsa (p. 531). From the Purbeck series have been obtained evidences of species of GoniophoUs distinct from the Green-sand and "Wealden kinds, as, for example, Petrosuclms (p, 636), Brachydedes (p. 643), Nannosuchus (p. 646), and Tlieriosuclius (p. 650).^ A Crocodilian modification of the Reptilian structure had, however, been attained as early in the Mesozoic periods as the Liassic period. Fam. — Protosuchii. Liassic Crocodiles of the genera of the extinct Protosuchian family are charac- terised, like most of their successors in subsequent Secondary periods, by biconcave 1 More complete evidences of structure tlaan those first acquired and on -wbicli the genera Getiosaurus, Foihilopleuron, were referred to the Crocodilian Order, have now shown that the species of those genera were more nearly allied to the Dinosauria. 17 130 BRITISH FOSSIL REPTILES. centrums, the primitive piscine form of vertebra under modifications, usually of a more consolidating kind, still prevailing. But the Protosuchians combined there- with jaws, longer and more slender than in existing Crocodiles and Alligators, armed with slender, conical, sharp-pointed and equal teeth, adapted like those of the existing Gavials, to the seizure and destruction of fishes. The Protosuchian species fall into two genera, characterised by the position and aspect of the external nostril, which aperture in one — called Steneosaurus — is situated a little behind and above the anterior termination of the upper jaw, in the other — called Teleosaurus — the nostril is at that end, or is ' terminal,' and looks more directly forward. Genus — Teleosaueus. Species — Teleosaurus Chapmanni, Plate 15 {Crocodilia), figs. 2, 2 a. The extinct reptile from which the characters of the genus Teleosaurus are derived, is one of the earliest of the evidences of ancient Reptilia which is recorded in a scientific publication. A brief description and figures of an incomplete skeleton found in the lias (alum schale) of the Yorkshire coast, about half a mile from Whitby, were published by Messrs. WooUer and Chapman, in two separate commu- nications, in the 50th volume of the ' Philosophical Transactions,' 1758 (Pt. 2, pi. xxii and xxx). Their figures exhibit a contorted and incomplete vertebral column, about 9 feet long, and a cranium, slightly displaced, 2 feet 9 inches in length. About ten vertebrae of the lumbar and sacral region of the trunk, and twelve vertebrae of the tail, remain in place ; the cervical, dorsal, and middle coccy- geal vertebrae were indicated only by their impressions, and these are fewer in number than the vertebrae in the existing Crocodiles. The skull is reversed, pre- senting its basal surface to view; the single occipital condyle, the zygomatic arches, terminated behind by the strong tympanic bones, and the large convex articular surface in each of these, for the lower jaw, placed in the same transverse line as the occipital condyle, are all recognisable. The skull appears to contract gradually to a pointed muzzle, but in reality to the base of a long and slender maxillary beak. In the remaining basal or posterior portions of the jaws the sockets of the teeth are seen separated by intervals of about 9 Unes ; in some of these there are pointed conical teeth which cross alternately those of the opposite jaw. The teeth are covered with polished enamel. Each of the vertebrae is 3 inches in length. Near the pelvic region, a portion of the shaft of the femur, including the head, was exposed, measuring between 3 and 4 inches in length. A few fragments of ribs were found near the dorsal vertebrae. The authors of the papers just analysed perceived suflBcient resem- LIASSIC CROCODILES. 131 blance between their fossil and the skeleton of the Crocodile to refer it to that family of reptiles ; but their figures and descriptions gave rise to various opinions respecting the affinities of the Whitby fossil in the writings of subsequent naturalists and anatomists. Camper, for example, pronounced it to be a whale, perhaps meaning a dolphin ; for, as Cuvier remarks, the presence of teeth in both jaws at once, proves the fossil not to belong to the Baljenffi, which have no teeth, nor to the Physeters, which have (conspicuous) teeth only in the lower jaw. Faujas adopted Camper's opinion, referring the fossil to the genus Physeter, and adding some reasons which are contradicted by the descriptions given by both Chapman and Wooller. Cuvier, in the first edition of his ' Ossemens Fossiles,' after refuting the opinion of Faujas, says, " La verite, ainsi que nous le verrons, est que c'etoit reellement un crocodile." The subsequent analysis, to which Cuvier here refers, led him in 1812 to the conclusion that it belonged to the genus of Crocodiles, and was most probably identical in species with the Crocodile of Honfleur. In 1836, however, when so many new and singular genera, allied to the Croco- dilian family, had been added to the catalogues of Palgeontology, chiefly by the labours and discoveries of English anatomists and geologists, Cuvier expresses his opinion on the fossil described by "Wooller and Chapman with more caution. He says, " II reste maintenant a savoir si c'est un crocodile, ou I'un de ces nouveaux genres decouverts dans les memes bancs. Les os des extremites y sont trop incomplets, et la tete n'y est pas represente avec assez de details pour decider la question ; mais les vertebres me paraissent plus longues, relativement a leur dia- metre, que dans les nouveaux genres, et plus semblables par ce caractere a celles des Crocodiles. Ceux qui retrouveront 1' original, s'il existe encore, pourront seuls nous apprendre si les autres caracteres repondent a celui-la." ^ A second specimen of a long and slender-nosed Crocodilian was obtained from the lias near Whitby, between Staiths and Eunswick, in the year 1791;^ and a more perfect skeleton was discovered in the alum shale of the lias formation at Saltwick, near Whitby, in 1824. Both these specimens so closely resemble the older fossil in all the points in which a comparison can be established, as to dissi- pate the remaining doubts as to the nature and affinities of the specimen from the same locality, described in the ' Philosophical Transactions ' for 1758. The skeleton, discovered in 1824, is figured in Young and Bird's ' Geological Survey of the Yorkshire Coast,' 2nd edit., 1828, pi. xvi, fig. J, p. 287, and in Dr. Buckland's ' Bridgewater Treatise,' vol. ii, pi. xxv. It is now preserved in the museum at Whitby, where I have closely examined it. In this specimen are preserved the cranium, wanting the snout, the whole vertebral column, the ribs, and the principal 1 " Eecherches sur les Ossemens Fossiles," torn. Seme, 2de partie, p. 114 ; 4to, 1824. ' See ' History of Whitby,' vol. ii, pp. 779, 780. 132 BRITISH FOSSIL REPTILES. parts of the four extremities, together with the dorsal and part of the ventral series of dermal bones. The entire length of the skeleton, follomng the curvature of the spine, is 15 feet 6 inches, to which may be added 2 feet 6 inches for the lost snout. The cranium posteriorly is broad, depressed, and square- shaped ; it begins to contract anterior to the orbits, and gradually assumes the form of the narrow depressed snout : the converging sides of the maxillee are con- cave outwardly. The zygomatic spaces are quadrilateral, longer than the upper temporal openings, and these are longer in the axis of the skull than transversely. The orbits are subcircular ; they look upwards and slightly outwards ; their margins are not raised, and their interspace is slightly concave. The parietal bone is relatively longer than in the Gavial, and sends up a longitudinal median crest, from the posterior part of which a strong process extends on each side outward, and curves slightly backward, parallel with the ex-occipitals, to join the mastoid and tympanic bones, the latter of which expands as it descends to form the joint for the lower jaw. Breadth of posterior part of skull Length of parietal crest .... Breadth of the interorbital space Antero-posterior diameter of the middle of tympanic pedicL Vertical diameter of orbit .... Antero-posterior of orbit .... Prom lower margin of orbit to alveolar border eet. In. Lines. 1 0 0 0 G 0 0 3 2 0 2 5 0 2 0 0 3 0 0 1 3 the entire length of the skull From these dimensions it may be calculated tha must have exceeded 4 feet 6 inches. The skull of one of the Caen Teleosauri measures 3 feet 4 inches, whence Cuvier calculates the entire length of the animal at near 15 feet. The Whitby Teleosaur agrees with the Caen species, and differs from the Gavial in the following particulars {Grocodilia, PI. 1, vol. ii). The anterior frontal is less extended upon the cheek ; the lacrymal is much more extended, and is larger at its base ; the malar bone is more slender. The post-frontal, which separates the temporal from the orbital cavities, is much longer and narrower. The parietal and occipital crests each form a thin trenchant plate, and are not flattened above. The mastoidean angle is not uninterruptedly united with the back part of the articular process of the tympanic, it is separated from it by a large depression, which is overarched by a trenchant crest belonging to the exoccipital. The mastoid has a concavity at its descending part, of which there is no trace in the Gavial. The indentation between the articular process of the tympanic and the tuberosity of the basioccipital is much smaller than in the Gavial, and the basilar tuberosity projects downwards in a less degree. The pterygoid ala is not expanded externally, as in all Crocodiles, but is contracted by a large fissure at the part where it goes to unite itself to the bone ; LTASSIC CROCODILES. 133 the orbital margin of tlie malar is not raised, and does not leave behind it a deep fissure as in the Gavial. The malar does not rise to join the postfrontal at the level of that bone, but this descends to join the malar at the external margin of the orbit. The vacuity between the orbit and mastotympanic is much elongated in the fossil, and occupies four fifths of the temporal fossa ; the anterior part of this fossa is narrow and acute. The columella or ossicle of the ear is cyhndrical, and much larger in proportion than in any known Crocodile or other reptile. Cuvier calculates the number of teeth in the Teleosaurus Gadomensis to be 180, viz. ffzl-f-. The Teleosaurus Ghapmanni has at least 140 teeth. The Gavial has IIZ, or 28-28- The teeth of the Whitby Teleosaur are as slender and sharp-pointed, but not so compressed, as in the Gavial ; they correspond with those of the Caen Teleosaur, and equally illustrate the dental characters usually attributed to the present extinct genus. The Whitby Teleosaur differs from the Caen Teleosaur, as does the Monheim Teleosam-,^ in having the upper temporal fossse longer in proportion, to their breadth ; but it differs from the Teleosaurs of both Caen and Monheim in the more equal size of the teeth, and from the Monheim species in the greater number of teeth, the Teleosmirus loriscus having at most f|z|-|=106. The median frontal in the Whitby Teleosaur is slightly concave, in the Caen species it is flat. The basi- occipital is perforated by the common terminal canal of the Eustachian tube close to the junction with the sphenoid, and on each side of the hole it expands into a rough tuberosity. The body of the sphenoid is compressed, characterised by two processes or narrow ridges, continued one from each side of the middle of the sphenoid obliquely backwards. The pterygoid bones are relatively smaller than in the Gavial. The palatine bones are more extended postei'iorly, and articulate with the transverse bones. The posterior apertures of the nasal canals are placed more forwards upon the base of the skull than in existing Crocodiles. Vertebral Golumn. — The number of vertebrse in the true Crocodiles of the present period rarely exceeds sixty, which is the number originally assigned by ^lian to the spinal column of the Crocodile of the Nile. Cuvier generally found 7 cervical, 12 dorsal, 5 lumbar, 2 sacral, and 34 caudal vertebrge. In the Crocodilus aaitus a thirteenth pair of ribs is occasionally developed, and, according to Plumier, it has two additional caudal vertebrge. The Alligator {Alligator Lucius) has 68 vertebrse, the additional ones being in the caudal region. The Gavial has 67 vertebra?, disposed as follows: — 7 cervical, 13 dorsal, 4 lumbar, 2 sacral, and 41 caudal vertebrae. The more perfect specimen in the Whitby Museum displays the number of the 1 Crocodilus priscus, Soemmerring. 134 BRITISH FOSSIL REPTILES. vertebra3 through the whole spinal column, and establishes another difference' between the Teleosaur and the Gavial, the former having a number of vertebrae- intermediate between the Crocodiles and Gavials, viz. 64, with a special peculiarity in the excess of costal vertebras, as the following formula indicates, viz. 7 cervical, 16 dorsal, 3 lumbar, 2 sacral, 36 caudal. In all sub-genera of existing Crocodiles, as in the extinct tertiary species, the hind surface of the vertebra is convex, the fore surface concave, except in the atlas and the two sacral vertebrae. Cuvier, who had the opportunity of seeing only the annular part (neurapo- physes) of the cervical vertebrse of the Caen Teleosaur, regrets his inabihty to state whether either of the articular extremities of the centrum were convex, or which of them.' The Whitby Teleosaur decides this question, and shows that both articiilar extremities of the vertebrae are slightly concave in the cervical as in the- rest of the vertebral series. The atlas in the Teleosaur corresponds essentially with that of the Crocodiles, as is shown by the three main component parts of this bone from a Whitby Teleosaur in Lord Bnniskillen's collection. The hypophysial centrum is a trans- verse quadrilateral piece, smooth and convex below, narrowing like an inverted wedge above, with six articular facets, viz. a concavity in front for the occipital condyle, a flat rougher surface on each side of the upper part for the attachment of the neurapophyses, a posterior facet for the anterior part of the true centrum, or ' odontoid element ' of the axis, and the small surface on each lateral, posterior, and inferior angle for the atlantal ribs. The neurapophyses are pyramidal pro- cesses, with their apices curved towards each other ; they are relatively smaller in proportion to the centrum than in the Crocodiles. The general anterior concavity for the reception of the occipital tubercle is formed at its circumference by the hypophysial centrum and neurapophyses of the atlas, and at its middle by the anterior detached odontoid, here evidently the homologue of the atlas in the Ichthyosaurus, the hypophysial centrum of the atlas in the Teleosaur representing the first inverted wedge-shaped bone in the Ichthyo- saur. The spine of the atlas is a large, strong, oblong piece, articulated with the neurapophyses of the atlas, and partly overlapping those of the axis. The cervical vertebrjB have strong transverse processes, a parapophysis developed from each side of the centrum, and a diapophysis from the base of each neurapophysis. The postzygapophyses look obliquely downward and outward, the prezygapophyses obliquely upward and inward. The spine is compressed, its base coequal with the whole antero-posterior extent of the nem^apophysis, its height equal to the distance from its base to the diapophysis ; it inclines slightly backward,. 1 ' Ossem. Fossiles,' 4to, 1824, torn, v, pt. ii, p. 137. LIASSIC CROCODILES. 135 ■and is rounded off at the summit. Tlie jDleurapophysis, or rib, is bifurcate at its vertebral end, the tubercle being as long as tlie head and neck ; its distal end is expanded into the liatcbet shape, the posterior angle being most produced, and overlapping the pleurapophysis of the next vertebra behind. The same mechanism for fixing and strengthening the neck thus existed for the advantage of the ancient marine Crocodiles, as we find in those of the existing epoch. In the dorsal region the ribs exchange the hatchet for the ordinary lengthened form, and soon begin to lose the head and neck, as in existing Crocodiles ; after the fifth they no longer articulate with the centrum, only with the diapophysis, which increases in antero-posterior extent and thickness, and presents an oblique notch at its anterior angle, for the reception of the tubercle, now the only head of the rib. The number of the dorsal ribs exceeds that of any existing Crocodilian, being, as above indicated, 16 pairs. The spinous process is proportionally strong ; in the "Whitby specimen it measures in most of the dorsal vertebra 2 inches in antero-posterior extent, and 7 lines in transverse diameter or thickness ; the height of these spines seems not to have much exceeded that of the cervical spines, but they are more truncated at the summit. A posterior dorsal or lumbar vertebra of a Teleosaur from the Whitby lias, in the collection of Mr. Eipley, corresponds with the vertebral characters of Teleo- saurus in the slight concavity and circular contour of the terminal articular surfaces of the body, and in the great antero-posterior extent of the spinous processes ; but that of the diapophysis does not exceed one half the length of the body of the vertebra, which is 2 inches 6 lines. This process is supported by two short, obtuse, sHghtly developed ridges, which rise from the upper part of the side of the body, as far apart as to include one third of the length of the body between them, and converge to the under part of the process ; a similar ridge extends from the upper part of the posterior end of the process obliquely backward to the base of the postzygapophysis. The neural arch is anchylosed to the centrum in this vertebra. The supporting buttresses of the diapophyses are not described by Cuvier in the dorsal vertebras of the Caen Teleosaur ; nor have I met with any dorsal or lumbar vertebi'se of the Whitby species, except the present, that was sufficiently perfect to exhibit this character ; it may, however, be constant and characteristic of the genus. It faintly indicates one of the most striking characters of the vertebrge of Streptosiyondylm. The anterior and posterior margins of the spinous processes are slightly excavated, and thus retain a character which is transitory in the Crocodile, and peculiar to an early period of its existence. The bodies of all the vertebrsB are compressed laterally, and concave antero- posteriorly at the sides ; but this character is more strongly marked in the anterior caudal vertebrje, which are flattened along the inferior surface; these vertebrse in the Whitby specimen were 2 inches 8 Lines in length. The diapophyses are 136 BRITISH "FOSSIL REPTILES. longer, but narrower antero-posteriorly than in tlie lumbar or dorsal vertebrae. The h^mapophyses are united at their peripheral end, forming chevron bones, but are detached at their central ends which are articulated, as in recent Crocodiles, with the interspaces of the vertebral centres. The caudal vertebrge progressively diminish in every diameter, save length, from the middle to near the end of the tail ; the terminal vertebrjB are shorter than the rest. The sternum and sternal ribs closely agree with the ordinary Crocodilian type. I have not yet seen a specimen of the abdominal sternal ribs. Pectoral extremities. — The scapula and coracoid resemble, in general form, those of the Crocodile, but are relatively smaller, in correspondence with the smaller size of the anterior extremities. The scajmla, for example, is only one third the length of the femur ; it is straighter than that of the Crocodile ; both, margins are nearly equally concave, instead of the anterior one being convex ; the humeral end is less expanded, and is more obliquely truncated. The coracoid is longer than the scapula, instead of being, as in the Crocodiles, shorter ; this prob- ably depends upon the breadth of the fore part of the body, which regulates the extent of the coracoid, while the proportions of the scapula more exclusively depend upon the development of the pectoral extremity. The coracoid of the Teleosaur differs also from that of the Crocodile in the greater expansion of its humeral end, the more transverse position of its sternal convex extremity, and a nearer approach, to parallelism in the direction of the two lateral margins. {Crocodilia, PL 1.) In the Whitby Teleosaur, discovered in 1824, the humerus of the right anterior extremity, and the humerus and bones of the fore-arm of the left (PI. 15), are preserved nearly in their proper relative positions. The humerus is shorter in proportion than in the Crocodiles, its length scarcely exceeds the antero-posterior diameter of two of the cervical vertebrae. The antibrachial bones are still more curtailed in their proportions ; the longest bone, or ulna, being not quite half the lenofth of the humerus. No portions of the carpal or other bones of the paddle are preserved, but the presence of the antibrachial bones, distinct from each other, and of the ordinary form and breadth at the distal end, forbid our supposing them to have been naturally deficient or of abortive proportions in the Teleosaurus. Admitting the humerus, radius and ulna to have existed for a purpose, that purjDose, we may conclude, from the modifications for an aquatic life in the rest of the skeleton, to have been the support and movement of a palmated manus ; an organ which would be of great use in turning and regulating the course of the swimmer and in bringing the long and slender snout, with the terminal nostrils, to the surface. The fore-paddles were doubtless much smaller than in ordinary Crocodiles, and this difference of proportion related both to the less frequent resorting of the LIASSIC CROCODILES. 137 Teleosaur to dry land, and to the light and slender character of its jaws and teeth with the consequent diminution of the weight of its head. (Crocodilia, PI. 1.) Pelvic extremity. —The pelvis of the Teleosaur was attached, as in the Crocodile, to the thickened and expanded transverse processes of two sacral vertebrse. These processes are stronger in the vertical direction, and intercept a relatively smaller and more regularly elliptical space than in the existing Crocodiles ; the anterior one appears not to have been so much expanded in the antero-posterior direction. The iliac bone seems to have been shorter in the antero-posterior diameter, but longer, as measured transversely to the axis of the trunk, and thus to have made a slight approach to its characteristic form in the Bnaliosaurs. Both the ischium and pubis are relatively more expanded than in the Gavial. The pelvic extremities are preserved in the Whitby specimen in nearly their true relative positions ; but the right is thrown directly over the left. The femur presents the usual form but is relatively more slender than in the existing Croco- dilians ; it is slightly twisted, and bent in two directions. Its proximal end is expanded, compressed with a regular convex curve, describing a semi-circle ; the trochanter is represented by a ridge which gradually subsides, and is lost upon the surface of the shaft. This is nearly cylindrical at the upper part, but is pro- duced at the anterior or convex side along the distal half in the form of an obtuse ridge. The condyles are very feebly indicated. In the Whitby specimen of 1824, The length of the femur is . The breadth of proximal end of ditto .... The diameter of middle of shaft ..... Both the tibia and fibula are subcompressed towards their distal end : the length of each bone is 8 inches. The shaft of the fibula is nearly as thick as that of the tibia. The bones of the leg of the Tehosaurus resemble those of Aelodon in their relative shortness as compared with the femur. In these, and probably in other ancient Crocodiles with biconcave vertebrae and marine habits, the tibia is little more than half the length of the femur ; while in recent Gavials it is two thirds that length. There are five tarsal bones, two in the proximal and three in the distal row, as in the Gavial ; but they are of more equal size ; the two proximal bones being by no means so disproportionately large. All the long bones have distinct medullary cavities, and these are even present in the metatarsals. In the Whitby specimen, The length of the middle metatarsal is . . . . . .6 inches. The breadth of its proximal end . . . . . .10 lines. The breadth of its distal end . . . . . . 6 ,, 18 Feet. lu. Lines. 1 3 3 0 2 10 0 1 4 138 BRITISH FOSSIL REPTILES. The ungual phalanges are dejoressed, smooth, and convex above, rounded at the end. Dermal armour. — The bony dermal scutes of the Teleosaur were regularly disposed like those of existing Crocodiles, in both longitudinal and transverse series ; the posterior margin of' one scute covered the base of the succeeding scute,^ and they slightly overlapped each other laterally. Cuvier states that one of the fossils of the Teleosaurus Cadomensis presents all those of one side in their natural situation, exhibiting, in the part of the body included between the first dorsal and the beginning of the tail, fifteen or sixteen transverse ro-^s, containing five scutes on each side ; so that there were at least ten longitudinal rows of these dermal bones. The scutes are arranged in the same manner and number, at least as regards the transverse rows, in the Whitby Teleosaur ; these rows being indicated by the large dorsal scutes still occupying their natural position in an uninterrupted line along the back ; they are twenty in number, and sixteen cover the vertebra included between the last cervical and first caudal (PL 15). The scutes of the Tel. Cliapnianni difi"er as much from those of the existing Gavials and Crocodiles as do those of the Tel. Cadomensis, being thicker, rectan- gular, and having the outer surface impressed with circular pits or indentations from 3 to 4 lines in diameter which are not confluent, but separated. The median dorsal scutes of the Whitby specimen are nearly square, having the longer diameter, about 3i inches across, placed transverse to the axis of the body, and with the outer margin slightly rounded. Each of these scutes is traversed, as in the Tel. ■priscus, by a longitudinal ridge, which is less developed than in the Gavials. The median dorsal scutes of the Tel. Cado- mensis and priscus appear to differ from those of the Tel. Ghapmanni iu being- more oblong transversely, and with the posterior and lateral margins rounded off". Cuvier does not allude to the carinated character of these plates in the Caen species. The lateral and ventral scutes of the Tel. Chapmanni are more perfect squares than those next the spine, but differ less in form and size from them than in the Caen Teleosaur. They are marked externally by the same impressed pattern, but are not carinated. The median abdominal scutes are not opposite but alter- nate ; their median margins are rounded off, or slightly angular ; and, while the anterior part of that margin is overlapped by the posterior half of the opposite scute, in advance, the posterior half overlaps the succeeding scutum of the opposite side. The verticillate cuirass of these ancient Crocodiles is thus securely braced round the trunk by this interlocking of the inferior extremities of each ring • Cuv., 1. c.,p. 279. LIASSIC CROCODILES. 139 of scutes, whilst the imbricated arrangement would allow of a certain sliding motion of the rings upon each other sufficient for the expansion of the chest in breathing. The scutes in the fine specimen in the Wliitby Museum measure about 5 lines in thickness, but are thinned off at the edge. Having now detailed the anatomical particulars which a study of the magnifi- cent and unique skeleton of the Teleosaurus, in the museum at Whiby, has enabled me to add to the previous descriptions, by Cuvier and other anatomists, of the osteological structure of this extinct Crocodilian genus, I next proceed to notice the principal examples of the same genus which are preserved in other collections of British Fossil Reptiles. The first of these is a fine skull of the same species of Teleosaurus, and from the same lias beds near Whitby in the museum of Mr. Ripley of that town : Feet. Inches. The length of the entire skull is . . . . . .29 From the angle to the beginning of the long symphysis of the lower j.iw 1 :i Breadth of the lower jaw at the posterior commencement of symphysis . . 0 2^ Breadth of the extremity of the lower jaw . . . .01 The extremity of the upper jaw well exhibits in this specimen the characteristic generic modification of its infundibuliform expansion, supporting the terminal nostrils, and resembling the extremity of the elephant's proboscis, wanting the digital process. This cranium also clearly exhibits the specific characters by which the Tel. rus Ghapmanni of the Yorkshire lias differs from the Tel. Cadomensis of the Caen oolite, viz. the greater antero-posterior extent of the upper temporal openings as compared with their transverse diameter in the Tel. Chapmanni ; the similar but slighter difference in the form of the orbits, the greater breadth of the interorbital space, which slightly exceeds the transverse diameter of the orbit instead of falling short of that diameter, as in the Tel. Cadomensis. A cranium of the Tel. Ghapmanni, in the museum of the Philosophical Institu- tion at York, and another in the museum at Scarborough, offer the same specific characters as the "Whitby specimens. In the Scarborough cranium the diameter of the orbit is 2 inches 3 lines, while that of the interorbital space is 2 inches 6 lines. In the museum of the Natural History Society at Lancaster there is a chain of five dorsal vertebrse of the Tel. Ghapmanni, from the Whitby lias, measur- ing 1 foot in length ; each vertebra is 2 inches 4 lines in length. A section of these vertebrge showed a small cavity in the centre of the cancellous structure of the body. Teleosa,urus Cadomensis. — Specimens of fragments of the jaw, teeth, and vertebra- 140 BRITISH FOSSIL REPTILES. of this species have been discovered in the Bath oohte at Bnslow, near Woodstock, and in the oolite at Stonesfield. Teleosaurus Cadomensis (var.). — Of this species, which is nearly allied to, if not identical with Cadomensis, I have examined a posterior cervical vertebra from the oolite near Chipping Norton, in the collection of Mr. Kingdon of that town- The sides of the centrum are less compressed than in the Tel. Ghapmanni, and the articular extremities have a more circular contour, the transverse exceeding the vertical diameter. There is no appearance of a ridge along the under surface : the transverse process of the centrum arises close to the neurapophysis. The length of this vertebra is Transverse diameter of centrum Vertical diameter of centrum Inch. Liues. 1 5 1 3 1 u- Species — Teleosaurus brevior {CrocodiUa, PI. 16). In Teleosaurus Ghapmanni (PI. 15) the skull, from the fore part of the orbit to the end of the snout, includes four lengths of the cranium behind that part of the orbit ; in Tel. brevior the part of the skull anterior to the orbit includes but two lengths and a half of the portion of the cranium behind that part of the orbit. The upper temporal apertures are subquadrate, the longitudinal diameter being the longest. The orbit is circular, but relatively smaller than in Teleosaurus Ghapmanni; as in that species the plane of its outlet is more oblique than in tertiary Crocodilia ; it combines a lateral as well as upper outlook. In tertiary species only the upper third of the cavity is seen in a profile view (compare fig. 1, PI. 16, with fig. 1, PI. 1 B, Grocodilus Hastingsioe). In advance of the orbits, in Tel. brevior, as in the type species {Tel. Ghapmanni), are the orifices representing the outer nostril in Enaliosauria. As in modern and tertiary Grocodilia, the hinder third of the mandible shows a large vacuity; but the splenial element, 31 (fig. 13, p. 97, vol. i), enters into the fore border of the vacuity in a larger proportion than in those later species. The massive paroccipito-mastoid productions are proportionally shorter (compare fig. 3, PI. 16, with fig. 2, PI. 1 a) ; the descending portions of the basioccipital are also shorter in proportion to their breadth, and are more definitely notched below ; the masto-parietals are longer, their free ends being more produced and slender. The proportion of the upper jaw retaining its smallest breadth is much less in length than in Tel. Ghapmanni. The breadth of the produced slender portions of both upper and lower jaws is relatively greater to their length than in Tel. Ghapmanni. The teeth are similar in shape, relative size, and number, to those of the type species, regard being had to the shorter jaws. LIASSIC CROCODILES. 141 The specimen figured is from the Lias of Wliitby, and is now in the British Museum. Species — Teleosaitrus latifrons, Owen (Grocodilia, PI. 17). The proportions of the cranial and facial parts of the skull in Teleosmirus Ghapinanni (PI. 15) are defined in the characters of the preceding species. In Tel. latifrons the part of the skull anterior to the orbit includes three times the length of the upper part of the cranium behind the fore part of the orbit. The entire length of the skull in the present species is about three times its greatest breadth, that is, taken across the back part of the temporal vacuities. In Tel. Chapmanni five times that breadth are included in the total length of the skull. Thus, in its general proportions Teleosaurus latifrons, like Tel. brevior, approaches near to the Geoffroyan genus Steneosaurus , but the terminal, almost vertical ex- ternal nostrils, the antorbital vacuities, with the relative slenderuess of the pro- duced upper and under jaws, are strictly teleosaurian. The occipital surface, viewed from above, describes a more regular concave curve than in Tel. Chapmanni or Tel. brevior, the frontal region between the orbits is broader, both absolutely and relatively, to the skull's length, suggesting the specific name. The parietal division of the temporal vacuities has been reduced by long exercise of the muscular masses therein lodged to a ridge, the walls of which diverge slightly at the fore part to give issue to the cerebral production traversing the " foramen pineale." The mastoid and squamosal supports of the tympanic joint are mai^ked respectively by the letters o and n, fig. 1. From the squamosals, the sides of the skull converge with a gradual and gentle curve to the maxillary elements of the upper jaw. The nasals terminate anteriorly, by their usual pointed inter-union, at the middle of the skull's length. The form of the transverse section of this part is added to figure 2 (palatal surface of skuU) ; also a similar section is shown, taken across the hinder-pointed ends of the premaxillaries. In advance of these the premaxillaries are slightly con- stricted laterally, but the fossil fails to give satisfactory indications of large inferior canines as the cause. Anterior to these the premaxillaries expand, and so surround the outer nostrils as to give them the almost vertical teleosaurian position. The palatal surface of the mandible and two sections are shown in fig. 4, PL 17. The specimen above described and figured is from the oolite of Northamp- tonshire. Species — Teleosaurus asthenodelrus, Owen. If the cranium of this Saurian should correspond with the characters of the 142 BRITISH rOSSIL REPTILES. genus which are exhibited by the vertebrjB and scutes here described, a distinct species is very evidently indicated by them, characterised by the smaller size of the cervical ribs, and the consequently weaker structure of the neck. In the Oxford Museum are preserved two cervical vei'tebraj and a dermal bone of this species, from the Kimmeridge clay at Shotover. The articular extremities and general form of the body of the vertebrjB accord witli the Steneosaurian type. Inches. Lines. The length of the centrum is . . . • .22 Vertical diameter of articular end ... ... 1 6 Transverse diameter of articular end . . . . . .15 Antero-po.sterior extent of lower transverse process . . . . 0 6 This process arises near the lower surface of the centrum, about half an inch from the anterior extremity of the bone. It is separated about the same distance from the upper transverse process, which is continued from the base of the neurapophysis ; both the supports of the cervical rib are one third smaller than the corresponding processes in the Teleosauri Ghapmanni and Gadomensis, and are less extended from the sides of the vertebra. The dermal scute is devoid of a ridge ; one half of the external surface is pitted with well-defined hemispherical depressions, separated from each other by about half their breadth, the smallest being nearest the margin ; the other half of the scute is smooth, and indicates that it was overlapped by the adjoining scute, according to the characteristic disposition of this fish-like covering of the present extinct marine g-enus of Crocodilians. &^ Species — Indetermina ta . In the Hunterian Collection are two entire dorsal vertebrae, with part of a third, fractured through the middle of the body, and displaying a small cancellated cavity filled with calcareous spar, as in the Teleosaurus Ghapmanni. These vertebras present the slightly cod cave articular extremities, and the other charac- ters of the genus Teleosaurus. The length of the centrum, measured along the under surface, is 2 inches G lines; vertical diameter of articular end 2 inches; transverse diameter 1 inch 10 lines ; transverse diameter of the middle of the body 1 inch. Both the inferior and lateral surfaces of the body are regularly concave, lengthwise; and smooth, except near the expanded articular extremities, where they are striated in the axis of the vertebra. The antero-posterior extent of the transverse process is 1 inch 6 Hues; that of the base of the spinous process 1 inch 9 lines. The transverse diameter of the spinal canal 7 lines ; its vertical diameter 4^ lines. OOLITIC CROCODILES. 143 These vertebrge are cemented together by a matrix, which closely resembles the gray Kimmeridge clay ; and a portion of a species of Peeten is attached, which is one of the characteristic fossils of the oolite group of secondary rocks, especially the Oxford clay. Genvf! — Steneosaurus, Geoffroy. CuviEE, after his instructive account of the remains of the " Gavial des carrieres depierre calcaire des environs de Caen,"^ to which Geoffroy St. Hilaire attached the name Teleosaurus, proceeds to describe the remains of another Gavial-like Crocodile from a different, Init oolitic, locality." On these subjects Geoffroy St. Hilaire remarks^ : " Je termme cette premiere lecture en prevenant que des objets representes en la planche vii des ' Osseviens Fossiles,' il u'y a I'applicables an teleosaurus que les sujets figures sous les nos. 1, 2, 3, 4, 5, 10, 11, 12, 14, et 17. Les autres objets (figs. 6, 7, 8, 9, etl3, 15, 16) venaient de plus loin de Quilly ; ils proviennent d'une autre espece, se rapportant a un autre genre que j'ai deja determine et nomme.* J'en traiterai ulterieurement sous la denomination de i>ieneosaurus." Of the figures illustrative of this genus are selected, for the present work, Cuvier's fig. 8 (PL 20, fig. 6), representing the palatal siirface of the upper jaw, one fourth of the natural size, and Cuvier's fig. 13 (PL 20, fig. 5), representing the upjjer surface of the fossil skull one twelfth of the natural size ; they represent the types of the genus. To the characters of Steneosaurus, thereon founded, I have been able to add those of the better preserved specimens, from the same geological zone, of a British locality, in Plate 18 {CrocodiUa). Geoffroy St. Hilaire recognised that the oolitic Crocodilian made a nearer approach to the Gavial than did the liassic Teleosaurus. The beak was relatively shorter, and the external nostril less terminal. He writes : " Nous verrons que ce genre est exactemeut intermediaire entre nos teleosaurus et le demembrement dvi grand genre crocodile, dont j'ai traite sous le nom de GavialisJ " It has not," he rightly remarks, " a skull so long and slender as in Teleosau^-us, but longer and slenderer than in Gavidlis." The transition towards modern or existing Crocodiles, which Geoffroy advocates in these ' Memoires,' is toward the 1 ' Ossemens Fossiles,' 4to., torn, v, pt. 2, p. 127. 2 lb., p. 134 3 " Divers Memoires sur de Grands Sauriens, &e.," ' Lu a I'Acad. Eoyale des Sciences,' le 4 octobre, 1830, p. 26. * ' Memoires du Musuum d'histoire iiat.,' torn, xii, 1825. 6 lb., p. 40. 144 BRITISH FOSSIL REPTILES. Gavialic modification, not going beyond in the direction of the short and broad-jawed forms, 3'et not quite reduced to the proportions of cranium and face exemplified in the more nearly allied long and narrow-jawed species at present existing in the great rivers of India; Geoffroy's remarks being exemplified by Cuvier's figures, representing the Steneosaurian characters, which I have introduced into PI. 20, (Grocodilia) of the present work. Species — Steneosaurus Geoffroyi, Owen, GrocodiUa (PI. 18, fig. 1). The mutilated cranial part of a skull (PI. 18, fig. 1), referable to this species, from the Great Oolite of Oxford, shows the following dimensions : Inches. Lines. Breadth of hinder or occipito-niastoid surface . . . . . 11 0 Height from the lower border of the occipital condyle to the parietal ridge . 4 8 Length of the temporal fossa . . . . . ..54 Breadth of ,, ,, ....... 5 0 From the intertemporal or parieto-frontal crest the sides slope at once, save at the fore part, where the crest expands to one inch in breadth with a slight superior convexity ; longitudinally the crest is slightly convex. The posterior boundary of the temporal fossa, formed by the parietal and mastoid bones (ib., ib., 7, 8) terminates above in a sharp ridge. The paroccipital (ib. 4) extends from the exoccipital outward to abut against the mastoid. The exoccipitals meet below at their junction with the basioccipital (fig. 2, 1) and exclude it from entering into the formation of the occipital foramen. In Teleosmirus, as in recent Crocodiles, the basioccipital contributes a small share to the lower border of the foramen. Above the suture, extending from the foramen outwards, the exoccipital is perforated by the precon- dyloid foramen, by the entojugular and the entocarotid foramina, the three being in a line extending downward and outward. The cranial canal is relatively smaller than in the Gavial and is subcylindrical. So much of the superior maxillary bone (fig. 1, 21) is preserved as shows sockets of 3 teeth in front of, and 27 teeth behind, a short diastema ; there is no groove along the mesial surface of the alveolar part. In the mandible the post- articular angle equals in extent the transverse diameter of the articular surface, approaching thus to pliosaurian proportions, whilst it is longer in the Gavial. The artictdar surface is convex in the middle, concave on each side, not uniformly concave as in the Gavial and modern Crocodiles ; the articular element extends more forward, and is broader on the inner side of the ramus. The depth of the ramus at the coronoid ridge is greater and the ridge itself is higher. There is no vacant interval between the angular and sub-angular elements. OOLITIC CROCODILES. 145 Species — Steneosaurus laticeps, Owen, CrococUlia (PI. 18, fig. 2). The breadth of the cranium in proportion to the length, measured from the occipital tubercle to the orbit, suggested the specific name. The oi'bits, like the temporal fossae, approach nearer to a circle in outer contour than in Sfen. Geoffroyi, and they have a less obliquely lateral aspect, approaching in that character nearer to the Teleosauri ; the facial part of the prefrontal, 14, terminates in a point, at the same transverse parallel as that part of the mid-frontal, n, and this part is relatively broader and shorter than in 8ten. Geoffroyi. The nasals, 15, become narrower, as they advance more rapidly than in that species, and terminate in a point at some distance from the external nostril which is bounded exclusively hj the premaxillaries, as in the type Steneosaur, which in this character adheres to the modern Gavial.^ Species — Steneosaurus temporalis, Owen, Crocodilia (PI. 19). This specific name was suggested by the great relative size of the temporal vacuities as indicated by their upper apertures (fig. 1, «, «)• The powerful biting muscles originating therein have reduced the cranial interspace to a ridge. The orbits seem to have been comparatively small; but the contour of their outlet had been mutilated in the clearance of the specimen from the oolitic free-stone (Bath) in which it was imbedded. The mastoidal process contributed to the concave mandibular joint is relatively broader than in the Teleosaur (PI. 16, fig. 3) ; the occipital condyle is relatively larger; and the descending part of the basi-occipital does not show the median notch. The length of the portion of man- dible (ib., fig. 4) from the angular process to the hindmost tooth-socket, indicates the proportion wanting in the interval between the cranial and orbital regions of the skull somewhat artificially joined in the Bath specimen. The outer nostril is bounded by the premaxillaries, the nasals terminating at some distance therefrom as in other Steneosaurs. The teeth, 24 on each side of the upper jaw (fig. 3), are relatively larger than in the preceding and type Steneosaurs ; and, with the general proportions of the facial part of the skull, indicate a transition to the platycoelian Crocodilia of later mesozoic formations. In the "Woodwardian Museum, Cambridge, is a portion of the skull of a Croco- 1 " Les 03 dii nez, k, k, sont bien eloignes d'aboutir a I'overture des narines," ' Oss. Toss.,' torn, cit., p. lOG. 19 146 BRITISH FOSSIL REPTILES. clilian, witli characters indicative of a Steneosaur; it is embedded in its matrix, and' includes the upper part of a cranium with much of the bony roof broken away ; this mutilation, however, exposes the cranial cavity, a cast of which gives a toler- ably good representation of the brain of the extinct reptile. The cerebral lobes, thus shown, have the smooth convexity of those of the Crocodile, and the similarly smooth optic lobes, but little less in size, are partially represented. The length of so much of the brain is 2 inches, the breadth of the cerebrum is 1-| inches ; the breadth of the skull is 6|- inches. The temporal fossae form wide ellipses 2 inches 9 lines in long diameter. From the back of the cranium to the beginning of the narrowing jaws is 8 inches; the length of the skeleton maybe estimated as about 18 feet. John Hunter had acquired for his series of evidences of extinct species — so remarkable at that period — a century ago — a portion of the mandible of a Steneosaur, including six inches extent of the hind part of the symphysis : the transverse diameter of the mandible at the junction of the rami is 4 inches, 3 lines ;. the hind surface of this union shows a deep transversely elliptical depression. Both upper and lower surfaces of this symphysial portion of jaw are flat, and the outer sides of the rami here are nearly flat and at right angles with the horizontal surfaces, the angles being rounded oS". The inferior flattened surface is impressed with small irregular longitudinal grooves, but not with pits or foramina. So much of the alveolar border is preserved in each ramus as measures 6-|- inches ; the inner border rises higher than the outer one, and this tract includes eight teeth of Steneosaurian character : the diameter of the circular base of the crown is from 4 to 5 lines. So much of the matrix as remains attached to this specimen resembles Oxford oolite. It is not without interest to see that the geologically younger Protosuchian shows narial and some slighter characters, by which it comes nearer to the modern gavialic modification of the order ; but the species of the genus next to be described, which have left their remains in still later mesozoic deposits, make a still nearer approach to the existing forms of CrocodiHan Reptiles. Gew(s— PLESIOSUCHUS, Owen,'' Species — FJestosuchus ManselUi," GrocodlUa (PI. 20, figs. 1 — 4). This genus and species exemplifies an interesting approach in an upper division 1 irXeo-ios, near; avyos, Crocodile (name of Egyptian species). ^ Steneosmiriis Manselii, Hulk, ' Quarterly Journal of the Geological Society,' vol. xxvi, p. 170, pi. is. KIMMERIDGIAN CROCODILES. 147 ^(Kimmeridge Clay) of tlie mesozoic series to the modern Crocodilia, as exemplified in the genera Crocodilus and Alligator, species of which have been restored from the older tertiaries (ante Vol. I, pp. 112 — 133 ; Crocodilia, Plates 1 — 4). The resemblance to these genera and difference from the genus Gavialis are manifested in the proportions of length and breadth of the antorbital part of the skull ; and the difference from the latter existing genus and the extinct Crocodilia {Teleosaurus and /Sieweosattnts), with gavialic proportions of the skull, is exemplified in the extension of the nasal bones (PI. 20, fig. 1, is) to the outer nostril, «, and is increased by the smaller proportional number of teeth in the upper jaw, 15, (PI. 20, fig. 2), than has been noted in any existing species of Alligator or Crocodile proper.^ Cuvier assigns to the latter genus 19 teeth on each side of the upper jaw f and to the Alligators 19 or 20 teeth in the upper jaw." In both genera, the teeth are signalised as of unequal size, " sont inegales " and this character, as illustrated in Cuvier's ' Planche I,' is more marked than it could have been in Plesiosuchus judging from the series of sockets, and the absence of the premaxillo-maxillary constrictions, shown by those recent species {Crocodilus vulgaris. Croc, acutus) which the fossil most resembled in the shape of the skull (PI. 20, fig. 1). The narial character which distinguishes the gavialic Teleosaurus and Steneo- saurns from the Crocodilian fossils is shown in Plate 1, fig. la, Gavialis, and fig. 2a Teleosaurus, which may be compared with PI. a, 2, fig. 1 [Crocodilus suchus) and PL 2a, fig. 1, Crocodilus champso'ides. The agreement in this respect with the latter group is exemplified in PI. 20, fig. 1, 15 n. But now comes an important, perhaps more important, character differentiating the present with all known mesozoic Crocodilia from all known tertiary and existing kinds, gavials not excepted. It is presented by the articular surfaces of the vertebral bodies or centrums ; these instead of articulating together by ball- and-socket joints, are joined by surfaces deviating little from flatness, and that in the dii'ection of concavity. Cuvier had recognised this character in the gavial- like fossils which he described and figured from Liassic and Oolitic formations ; but he was not moved thereby to frame for them generic names. Of the vertebral body in his ' Gavial de Caen' (Teleosaurus, Geoffroy) Cuvier states : — " II a ses deux faces tres-legerement concaves, et son milieu retreci." (Op. cit., p. 137), and he remarks : — " C'est la, comme on voit, un caractere fort different de celui des crocodiles vivans, oil toutes les faces posterieures sont tres- convexes et les anterieures tres concaves" (lb., ib.) 1 " Les CROCODILES — proprement dits — out quinze dents de chaque cote en bas, dix-neuf en haut," ' Ossemens Possiles,' torn, v, pt. ii, 4to, p. 31. - " Les Caiuans {AlUgatof) ont, au moins, dix-neuf et quelquefois jusqu'a vingt-deux de cliaque •cote en bas ; au moins, dix-neuf, et souvent vingt en haut," lb., p. 30. 148 BRITISH FOSSIL REPTILES. Fortunately portions of two vertebrae are included in the mass of matrix still adherent to the skull of Plesiosuchus. The first is a centrum of a mid- or post-dorsal vertebra (PL 20, fig. 3) lodged in that part which obscures the base of the skull j it gives a length of oj inches ; a vertical (posterior) breadth of 3 inches ; the transverse breadth seems not to have exceeded 2 inches at the middle of the centrum. The preserved portion of the base of the neural arch shows that the parapophysis had disappeared or risen into confluence with the diapophysis, indicating the region of the spine from which this vertebra has been derived. The character of the terminal articular surfaces of the centrum removes this form of mesozoic Crocodile from the tertiary or neozoic genera. Both surfaces are slightly concave, the fore one nearly approaching to flatness. The degree in which the skull conforms to the proportions characteristic of Cuvier's genus Grocodibis proper,'^ difi"erentiates it from the gavial-like genera Teleosaurus and Steueosaunis, and the narial character above referred to more decisively separates it from the latter, with which it has been confounded. I am, therefore, constrained to append a name to the extinct genus which the Kim- meridgian specimen represents, retaining the specific name applied to it by Mr. J. "W. Hulke. I regret that having to record it, agreeably to the wish of my friend Mr. Mansell Pleydell, in the concluding volume of the present Work, with con- sequent unavoidable delay, gave the opportunity of a feUow-labom^er in Reptilian PaliBontology to make the specimen known, and, in anticipation, to append to it a generic name, which if applicable, I should have felt bound to retain." A second vertebra includes, with part of the centrum, the neural arch and spine. The latter shows its obtuse free termination. The total height of this vertebra is 9 inches ; that of the neural arch, spine inclusive, is 6^ inches ; the neural spine along the fore part is 4 inches ; along the hind jiart from the base of the postzygapophyses it measures 3 inches ; the thickness of the spine is 1 inch ; the broad base of the broken diapophysis is 2 inch.es. This process projects from near the base of the neural spine. The vertebral characters bespeak a platycoelian Crocodile of great power, in concordance with the formidable array of teeth. The length of the maxillo-premaxillary alveolar tract (PL 20, fig. 2) is 1 foot, 8^ inches ; it includes 16 teeth on each side, of which 3 are supported by the pre- maxillary, and the rest, after an interval of 9 lines, by the maxillary bones. Of ' As exemplified in Crocodilus acuius, torn, cit., p. pi. i, fig. 3. 2 " A closer examination lately made by Mr. Davies, sen., of the fossils presented to tLe British Museum last year by J. C. Mansel, Esq., has led to the identification of a large Crocodilian head — this head had been previously put aside as Pliosaurian " (Hulke, ' Quarterly Journal of the Geological Society of London,' vol. xxvi (1S70), p. 167). That this collection included Pliosaurian remains from the same Kimmeridgian bed is true ; but that the Crocodilian head was referred by any authority of mine to the Sauropterygia is incorrect. KIMMERIDGIAN CROCODILES. 149 tlie premaxillary teeth the third is rather larger than the rest ; the maxillary series begins by one somewhat smaller, the rest are of larger size, and maintain it to the thirteenth, the remaining two slightly decreasing ; but a general uniformity prevails throughout the series. The base of the ninth tooth has a diameter of 1 inch, 2 lines, with a circular transverse section; that of the second (pre- maxillary) tooth gives 10 lines in the same direction. The upper jaw, with a basal breadth of 7 inches, gradually and uninterruptedly narrows to 2^ inches across the foremost pair of sockets. The palatal interspace between these sockets does not exceed that which separates the socket of the first from that of the second tooth. This interspace seems not to have exceeded 4 lines at the alveolar outlet ; the interspaces of the maxillary alveoli are so much less, that most of the teeth of the same series between the two extremes must have been nearly, if not quite, in contact basally. The breadth of the palate between the third pair of teeth is 2 inches ; between the last pair it is 8 inches. The mid- line of the mouth-roof is slightly produced. Portions of the broken off crowns of the teeth show the two-ridged Crocodilian character, with a smooth intervening enamel. Two of these detached crowns (ib., fig. 4) seem to indicate the extent to which the upper and lower teeth over- lapped in the closed mouth; of these the best preserved gives a length of 4|- inches. A transverse fracture of a tooth-crown, where its diameter had fallen to 7 lines, shows a pulp-cavity diminished to 3 lines. The upper aperture of the temporal fossa is six and a half inches in transverse, and seven inches in fore-and-aft, diameters. The upper (post-fronto-mastoid) boundary curves outwardly, with a thick convex surface. The mandibular muscle, which it included, had a size conformably with the close-set series of large teeth, which it worked. The frontal bones, ii, converge to a point, ten inches distant from the premaxillary apex of the skull. The nasal bones, is, gradually narrow to a point penetrating the hind border of the nostril, n. This opening is ovoid, three inches and a half in length, three inches in greatest breadth. The premaxillaries meet and join an inch and a half anterior to the horizontal nostril.^ The transition to tertiary and existing Crocodiles is manifested by the propor- tions of the skull and of the teeth ; but these, in the degi'ee of general equality of size, are gavial-like, while in relative size and paucity of number they show the Crocodilian character in excess. There is no trace of an alveolar pit in the upper jaw for the reception of a lower canine, as in the Alligators, nor of any lateral notch for such a tooth as in 1 " Les intermasillaires, a, a (figs. 1, 2, 3), entourent les narines externes, excepte un endroit fort etroit ou la pointe des os nasaux, Tc, k, se place entre eux." " Determination des OS de la tfite dans les Crocodiles proprement dits." ' Ossem. Fossiles,' v, pt. ii, pp. C9, 71. 150 BRITISH FOSSIL REPTILES. the Crocodiles, The general equality of size in the tooth-crowns seems a remnant of the earlier mesozoic dental character ; but the number of teeth is even less than in any known modern Crocodile or Alligator. Of the latter the species which I have examined show eighteen teeth on each side of the upper, and as many teeth on each side of the lower jaw. Of CrococUlus (Cuv.), I have not seen any species with fewer than eighteen teeth on each side of the upper, and fifteen on each side of the lower jaw. In no instance is so small a number recorded as sixteen teeth on each side of the upper jaw, characteristic, as far as the present unique skull shows, of the genus Plesiosuchus. Existing Crocodiles differ from Lizards in the position and relative size of the opening, by which the breathing-passage from the nostril communicates with the roof of the mouth. The bony palatonarial aperture is single, small, and placed far back, very near the basi-occipital condyle. Those examples of extinct Liassic and Oolitic Crocodiles in which, from parts of the fossil skull, usually fragmentary, a judgment could be formed of this character, had led me to the conclusion that both in size and less retral position the palatonares showed more resemblance to that in Lacertians, and thus gave indications of a more generalised Saurian structure. CuviEB, in his description of a fossil skeleton of a Crocodile {Steneosaurus, Greof .) from the Caen Oolite, recognised the palatonaris as a large aperture far in advance of the position it presents in existing Crocodiles.^ This distinctive character was, however, called in question by later Palfeonto- logists. In the ' Abhandlungen liber die Gavial-artigen Reptilien des lias- formation,' fol. 1841, by Professors Beonn and Kaup ; it is argued at length (pp. 12, 16, 24) that the posterior median foramen is the true hinder or palatal aperture of the nosti'ils ; and a letter from De Blainville is cited by those authors in support of their view, in which Cuvier's determination, that it was ' an arterial foramen,' is rejected, antl Professor Beonn's opinion is stated to be completely confirmed by the appearances in the fossil skull of a Teleosaurns from Caen. After an examination of the original specimen, described by ChajDman, (' Phil. Trans.,' vol. 1), and now in the Whitby Museum, I proceeded to compare the foramina in question with those, the true nature of which might be determined by anatomical investigation of their relations and functions in existing Crocodiles and Gavials. The results of these dissections and injections were communicated to the Royal Society,^ and they demonstrated the accuracy of Cuvier's ascription of the characters of a palatal nostril to the larger and more advanced vacuity in the bony palate ; but they brought to light a more complex structure of the communications ' " La fosse nasale posterieure est tres grande, et biea eloiguee de ne s'ouvrir que vers I'extremite de la face basilaire, ou sont dans les crocodiles ordinaires les arriere-narines tres-peu avant le trou des arteres . ." ' Ossetnens Fossiles,' torn, v, pt. ii, p. 133. 2 ' Philos. Trans.,' 1850, p. 522, plates .10— 42. KIMMERIDGIAN CROCODILES. 151 between the organ of hearing and the palate, and showed that the ' foramen ' which Olivier had described as giving passage to an artery was the medial outlet of an eustachian canal, one by which the atmospheric air could pass to the tympanic cavity. The foregoing details of the osteological and dental characters of the extinct Crocodilian reptiles of the secondary or mesozoic formations, as well as those of the tertiary Crocodiles in the first volume of the present work, testify how little was left by Baron Cuvier to be added by his disciples and successors. The combination of secondary and tertiaiy modifications in the structure of skull and teeth, in the Kimmeridgian PlesiosucJms, will probably be esteemed as the chief addition to this chapter of Extinct BeptUia. It owes to Geoffrey St. Hilaire the invention and application to genera, the chai^acters of which are exclusively due to Cuvier, of names which the founder of paleeontological science cared not to give. In my PL XX the two figures from the PI. VII of the " Ossemens Fossiles " are repeated, which Geoffroy selected as characteristic of the genus he proposed to term Steneosaurus. But the gist of the ' Memoires ' of 1825, as of those of 1830, was to interpret the Cuvierian facts according to the Lamarckian evolutional hypothesis of Species, which Geoffroy had adopted in the following terms : " La formation successive et leur evolution dans le cours des figes." He then proceeds : — " Je montrerai des formes remplacees insensiblement par d'autres, qui n'auraient pu s'accommoder de I'ancien ordre des choses;" in other words, the battle of life was against them. Reflecting on the Palseontologist who guided his course in the science, "par de faits jwsitifs,"^ Geoffroy afl&rms, "qu'il renonce a ce qu'il y a de plus vif, de plus enivrant, et de plus profondement philosophique dans la vie des sciences."" " Les modifications insensibles d'un siecle a uu autre finissent par s'ajouter et se reunissent en une somme quelconque. Si ces modifications amenent des effets nuisables, les animaux qui les eprouvent cessent d'exister, pour etre remplaces par d'autres, avec des formes un peu changees, et changees a la convenauce des circonstances." See also the ' Section ' entitled ' Le degre d'influence du Monde ambiant pour modifier les formes animales,' 4eme Memoire, p. 79. Under this conviction Geofi'roy rejoiced to see the transitional step, though short, which the extinct oolitic Crocodile, of " Quilly," his Steneosaurus, made toward the modern Gavial. Still more exultant would have been his reception of the form here described of an advance made in a secondary formation, nearer our times, beyond the gangetic long-beaked Crocodile to the shorter and broader cranial characters of the more numerous and widely distributed " Emydosaurians," representing the existing genera Crocodllus and Alligator. 1 ' Divers Memoires,' &c., p. 137. - lb., ib. 152 BRITISH FOSSIL REPTILES. CHAPTER V. Oeder— SAUROPTERYGIA, Owen. Genus — Pliosaueus, Owen. In the genus of Sauropterygia called Plesiosaurus (ante, p. 1) the cervical vertebra exceed in number and the neck in length those of other Saurian genera, and the skull, so supported, is small in proportion. But, among the species defined and described in the present volume (pp. 1 — 34), a certain range of variety is exemplified in these respects. In Plesiosaums Iwmalospondijlus, for example (pp. 12 — 20, Tab. V), the cervical vertebrse are thirty-eight in number, and together include seven lengths of the skull. In Plesiosaurus rostratus (pp. 20 — 33, Tab. IX) the same vertebras are but twenty-four in numljer, and together equal only one and a half the length of the skull they support. The direction of specific modification here indicated gave warning of the secondary importance of the cervical character, and at the same time suggested a possibility of a still larger head with concomitantly shorter neck having been engrafted on an essentially sauropterygian type of structure.' The discovery of certain fossil teeth which combined sauropterygian characters with modified shape, and a size much exceeding that of any of those in the then known Plesiosauri, led me to make them known as indicative of a distinct genus under the name Pliosaurus." These teeth {Saiiropterngia, PI, 33, vol. iv) have a crown thicker in proportion to their length and three-sided in shape, showing a subtriedral transverse section, with one side flattened (ib., fig. 3), and bounded by prominent ridges from the more convex sides, which are rounded oflf into each other, and alone show the well- defined longitudinal ridges of the enamel (ib., figs. 1 and 2). The cervical vertebra (Sauropterygia, PL 18) are so compressed from before backward (fig. 1) as to approach the ichthyosaurian type (pp. 45, 85, Tab. xvii, xxix), but the articular surfaces are almost flat, (fig. 2) ; and though I have found as many as twelve in one individual, they are so compressed as to cause a very short neck to intervene between a large head and a massive trunk, in which the dorsaUvertebrge resume the ordinary plesiosaurian proportions. 1 See " Eemarks on this subject," in connection with a proposed Order called Macrotraclielen, by V. Meyer, in my ' Palaeontology,' 8vo, 1851, pp. 252, 253. - ' Report on British Fossil Eeptilia,' Part ii, 1S41, in " Eeports of the British Association for the Advancement of Science " for that year. KIMMERIDGIAN PLIOSAURS. 153 A cervical vertebra {Sauropterygia, PI. 18) of aPliosaiirus, fi^om the Eammeridge Clay of Foxcombe Hill, near Oxford, measures, for example, in breadth six inches ; in depth, or vertical diameter, five inches ; while in length, or the diameter corre- sponding with the axis of the animal's body, or of its vertebral column, it measures only two inches and a half. Nevertheless, with these ichthyosaurian proportions is associated an essentially plesiosaurian type of structure. The lower surface of the cervical centrum (ib., fig. 2) shows the pair of vascular foramina; the terminal articular surfaces are flat or very sHghtly concave : the cervical rib was ligamentously tied, in some species, to two processes, the di- and par- apophyses (fig. 1), occupying two thirds of the fore-and-aft extent of the side of the centrum, slightly projecting beyond the surface, and divided by a deep linear fissure. I have rarely seen an instance in which the neurapophyses were anchylosed to the centrum, and never one with the pleurapophyses so attached. At the trunk end of the series the costal processes begin to climb, as in Plesiosaurus, upon the neurapophyses, — the diapophysis growing at the expense of the parapophysis, until the rib becomes supported, in the dorsal I'egion, upon a single strong and prominent process : this is subdepressed, with an oval transverse section, which is rather sharp at the anterior margin. The vertebral centrums begin to gain in length as the costal processes rise in position, and those of the dorsal region have attained to quite plesiosaurian proportions. Throughout the rest of the column the vertebrje closely repeat the plesiosaurian characters on a large scale. The sides, or non-articular surface of the centrum, are rugous near the articular ends, elsewhere smooth, and in the dorsal region longitudinally concave. In the caudal vertebras the costal process is undivided, prominent, with a vertically elliptical section, continuous with the neurapophysial surface at the base of the tail : the lower surface of the centrum is square-shaped and nearly flat : its angles are marked by the hypapophysial surfaces, of which the anterior pair is usually the largest. The generic character derived from the organs of locomotion is the apparent absence of the antibrachial and cnemial bones, which seem to be represented by a jDroximal row of three large " carpal " and " tarsal " ossicles. On the homology of these I shall offer remarks in the sequel. As to the history of the present genus, I may briefly state that in a ' Pieport on British Fossil Reptiles,' communicated to the Meeting of the British Association for the Advancement of Sciences, held in 1839, and jarinted in the volume of ' Reports ' for that year,^ I described certain fossils, from which were deduced the two species of Plesiosaurus, called "• grmuUs" (p. 83), and " trochanterius" (p. 85). In my second Report on the same class of fossils communicated to the Association in 1841, I pointed out (p. 60) the characters by which these two species departed 1 ' Eeport of Brit. Assoc.,' Svo, pp. S3, 86, 1839. 20 154 BRITISH FOSSIL REPTILES. so far from the type-characters of Plesiosaurus as to merit being placed in a distinct genus or subgenus, for which I proposed the name of Pliosaimis ; admitting at the same time in reference to the two species, that " subsequent discoveries and observations were needed to supply distinct and recognisable characters for them " — " the two forms of femora, on which they were founded, not having then been found so associated with vertebrae and other bones as to aid in their definition." ^ I am now enabled to describe and figure specimens, among those that have subsequently come under my notice, which afford good grounds for the acceptance of the two species, and for the addition of a third to the genus Pliosaurus. It may seem strange that jaws which have lost all their teeth should yield new characters derivable from the number, proportions, and disposition of such organs ; but herein a Paleontologist's mode of work is like that of Antiquaries of another order, who read inscriptions on Roman buildings by the nail-marks when the letters themselves have been wrenched off for the sake of the metal. Species — Pliosaurus grancUs, Owen (Sauropferygia), Plate 19, figs. 1 and 2. PLESiosATJErs OEANBis, Ow. Eepopfc on British Fossil Eeptiles, 8vo, p. 83, 1839. PLiosAUErs BEACHTDEiEUs, Ow. Odontography, 4to, p. 283 (?), 1840. The most complete example of the skull of a Pliosaur which has come under my observation was disinterred from the Kimmeridge Clay, at Kimmeridge, Dorsetshire, under the superintendence of J. 0. Mansel-Pleydell, Esq., F.G.S., of Longthorns, in that county. This skull is also the largest of such specimens hitherto found ; and, since the matris has been removed, it has yielded the most instructive chai^acters of cranial structm^e and dentition. Originally sent to me with the skull of Plesiosuchus (p. 147, pi. 20) for determination and description, both specimens have since been presented by their discoverer to the British Museum. The same liberal donor has subsequently enriched the National Collec- tion by a lower jaw and part of the cranium, with evidence of the locomotive organs of the Pliosaurus trochanter ius. I shall premise to the descriptions some of the dimensions of remains of both specimens. 1 lb. (Second Keport), p. 54, 1841. KIMMERIDGIAN PLIOSAURS. 155 Table op Admeasurements. PUosaurus grandis. PUosaurus trochanterius. Feet. Inches. Lines. Feet. Inches. Lines. Lengtb of the mandible in a straight line from the fore end of the symphysis to the angle of the jaw lb., following the curve on the outer side of the ramus . Length of the alveolar series following the curve Length of the "symphysis mandibulse" Greatest breadth of ditto .... Breadth at posterior part of ditto Depth of, at the same part .... Depth of ramus anterior to articular surface Breadth of mandible at the hind end of the alveolar series ....... Length of cranium from the occipital condyle to the end of muzzle 1 ..... G-reatest breadth of premaxillaries, viz. across the fourth pair of alveoli ..... Breadth behind the iifth pair of alveoli Breadth of cranium behind the alveolar series . Length of palato-nares ..... Breadth of ditto ..... Distance from their back part to end of occipital condyle 5 5 3 1 1 4 o 8 11 7 1 7 6 5 4 11 9 G 5 1 5 1 6 6 3 9 6 3 9 G 8 5 4 4 1 1 4 5 9 4 7 7 3 4 8 6 6 6 4 4 9 9 8 In PUosaurus grandis the number of alveoli on each side of the uppei' jaw is twenty-seven or twenty-eight. The first pair (PL 19, fig. 2) are terminal and approximate ; the outlet of each measures 1 inch 3 lines in long diameter, which lies in the axis of the jaw : the second alveolus with au outlet 1 inch 9 lines in long diameter, which is transverse to the jaw's axis, is divided by a partition of 4 lines breadth from the first : the third socket is divided by a partition half an inch in breadth, from the second; its outlet is circular, and 2 inches in diameter : the fourth socket is of similar size, and at rather less distance from the third : the fifth socket is less in transverse diameter than the third, but is equal in fore-and-aft diameter to the fourth, from which it stands 9 lines apart. These five pairs of alveoli are in the premaxillary bones (22), and occupy the whole of their alveolar extent. An interval of rather more than two inches inter- venes between the last premaxillary and the first maxillary alveolus, which is the sixth of the series ; and this interval is traversed obliquely by the maxillo- premaxillary suture (PI. 19, fig. 2). The maxillary alveoli have partition walls of about 4 lines in thickness at their fi'ee border ; but these become thinner as the teeth decrease in size in the hinder third part of the series. The alveoli increase in size from the first to the fourth maxillary tooth (ninth of the dental series) ; the longest diameter of the aperture of this socket is 3 inches : thence the alveoli 156 BRITISH FOSSIL REPTILES. gradually decrease in size to a diameter of half an inch. The form of the alveolar aperture is, for the most part, a full oval, nearly circular, with the long diameter inclining more or less transversely. The margins of the larger maxillary alveoli are the most prominent. The entire alveolar series describes longitudinally and horizontally a gently undulated course, the premaxillary series forming a slight convexity, and the larger maxillary alveoli a similar convexity, outward. Longitudinally and vertically the alveolar border is almost straight as far as the seventeenth tooth, and then gently bends upwards to the hind end of the series. A groove, deepening into fossas answering in number to the alveoli, extends along the inner side of each premaxillary series. This groove is interrupted at the diastema between the premaxillary and maxillary alveoli : it recommences at the inner side of the maxillary series, also deepening into pits opposite the inner and back part of the alveoli, and continues, though feebly indicated, along the hinder third of the alveolar series. The bony palate is entire, save at the palato-nares (ib., fig. 2, r, r) ; but on the inner side of the twelfth socket, counting backward, on each side, there is a nervo- vascular foramen terminating a canal in the upper jaw, directed obliquely down- ward and forward : the foramen is elliptical, an inch in diameter ; a shallow channel extends a few inches in advance of its outlet, and three or four similar but smaller foramina succeed each other anteriorly near the inner wall of the internal alveolar groove, leading to a linear channel 7 inches long, which, with its fellow on the opposite side, defines the base of a median longitudinal ridge of the bony palate between the first three pairs of maxillary alveoli, which ridge is transversely convex, and about an inch in breadth. As the bony palate expands in breadth, behind the nervo-vascular foramina, it presents trans- versely a broader median convexity, bounded by lateral shallow concavities. On the transverse line, between the sixteenth pair of alveoli, are the anterior ends of the palatine bones, which are divided by a median suture (20, 20). The major part of the palato-nares are bounded by the pterygoids (21, 24), which extend backward to the base of the occipital condyle, underlapping the basi-sphenoid and basi-occipital, developing ridges which project below the level of the lateral parts of the fossa, and converging to meet behind the area, including the posterior nostrils. External to these ridges the pterygoids diverge to abut against the tympanic pedicles.^ The mesial border of an ectopterygoid is preserved at 25, fig. 1 . The number of alveoli in each ramus of the mandible (PI. 19, fig. 1) is twenty- five or twenty-six. The five alveoli corresponding with the premaxillary sockets in the upper jaw are the largest. They are separated by similar intervals. ' The tympanic articulations of the lower jaw, which extend the cranium beyond the condyle, are broken off. KIMMERIDGIAN PLIOSAURS. 157 Between the fifth and the sixth alveolus is a diastema of about 8 lines ; the long diameter of the sixth alveolus is 1 inch 10 lines. An interval of 5 lines divides it from the seventh socket. The succeeding ones are closer together : they gradually increase in size to the twelfth or thirteenth, but do not obtain the size of those opposed to them above ; they then gradually decrease in size and depth to a diameter of about half an inch. The summits of crowns of successional teeth protrude from fossEe at the inner and back part of the anterior alveoli. The crown of a more advanced successional tooth projects into the bottom of the socket of the third and fifth of the symphysial series : these teeth show the characters of the genus Pliosaimis. The inter-alveolar part of the "symphysis mandibulfe" forms a median longitudinal rising, less convex or ridge-like than the one on the palate above. FossEe are discernible on the inner side of the mandibular alveoli, but less marked in the upper jaw. The apex of a successional tooth appears in two of these pits. On the inner side of the posterior third of the mandibular ramus there is a wide and deep channel between the surangular (29) and angular (30) elements ; and this groove is continued forward indicative of the upper border of the splenial (31) which extends along the inner side of the lower half of the dentary nearly to the symphysis. The articular surface of the mandible (29), 7 inches is transverse, and 5 inches in antero-posterior extent, is slightly concave transversely at the inner three fourths of its extent, and then gently convex at the outer fourth ; it is more concave from before backwards in the major part of its extent, but the peripheral boundary is not entire. The sauropterygian affinities, as contradistinguished from the ichthypterygian, are exemplified in the more complete and separate sockets of all the teeth, and in the smaller projDortion contributed by the premaxillaries to their support and to the formation of the upper jaw. The palato-nares of PUosaurus are more linear and approximate than in the species of Plesiosanrus {PL Haivhinsii, PL XVI ; and PI. rostraius, PI. XIII, r, r) in which they have been observed. Species — Pliosmtrus brachydeirus, Owen. In the Museum of Geology at Oxford are considerable proportions of the upper and lower jaws of a PUosaurus from the Kimmeridge Clay at Market-Raisin.^ The teeth, in proportions and arrangement, correspond so closely with those in the specimen above described as induced me to surmise that they might belong to the same species. The following are the difierences which I have noted between 1 ' Second Eeport on British Fossil Eeptiles,' " Eeport of British Association," p. 61, 1841. 158 BRITISH FOSSIL REPTILES. them : tte widest diastema divides the fourth upper tooth from the fifth in the Oxford specimen, not the fifth from the sixth : the maxillo-premaxillary suture with the lateral compression at this interval, is as in the British Museum specimen. If the pair of small anterior sockets and teeth are wanting, either through age or accident, in the Oxford specimen the difference noted would be accounted for. It may be remarked that the number of alveoli — twenty-six — on the least imperfect side of the upper jaw is the same in both skulls, and in both a small part of the series is wanting posteriorly. In both the premaxillary part of the jaw containing four pairs of large teeth is slightly expanded. In the maxillary part of the Oxford specimen the teeth increase in size to the sixth ; in the British Museum specimen to the fifth ; beyond which they gradually diminish. The length of the best-preserved alveolar series is 3 feet in the Oxford specimen, and 3 feet 7 inches in that in the British Museum. In the mandible from Market-Raisin there are thirty- five sockets in each side ; in that from Kimmeridge there are only twenty ; but as neither specimens have the alveolar series quite complete, I do not feel that there is sufficient ground to reject the hy^Dothesis of individual variety. In all the essential characters, including length of symphysis mandibulte, the Market-Raisin skull agrees with that in the Kimmeridge example of PUosaurus grandls, and differs from that of Pliosmirus trochanterius, next to be described. If, however, the minor differ- ences which have been noted between the Oxford specimen and that figured in PL 19, figs. 1 and 2, should prove to be constant, the specific name " brachydeirus," by which I originally indicated Dr. Buckland's magnificent specimen^ from Market- Raisin, might be retained for it. Species — PUosaurus troclianferms, Owen (Sauropterygia) Plate 19, figs. 3, 4, 5. Plesiosauetis TEOCHANTEEirs, Ow. Eeport on British Possil Eeptiles, Sto, p. 85, 1839. In the work above cited the specific character of the fossil Reptile in question was indicated by modifications of the femur ; but the chief distinction between PUosaurus trochanterius and PL grandis is conspicuous in the greater relative extent of the symphysis mandibuljs in the former, and in the greater proportion of the dental series lodged in that part of the lower jaw. This character is exemplified in the fourth admeasurement in the " Table," p. 156, and in PI. 19, fig. 4, as compared with PI. 19, fig. 1. The surangular developes in PUosaurus trochanterius [fig. 29'] a low but well- 1 ' Odontography,' p. 283. KIMMERIDGIAN PLIOSAURS. 159 marked angular coronoid process. Anterior to tliis tlie upper border of the mandible becomes tliick and transversely convex; and, an inch below tbe border, the outer side of the ramus is impressed by a wide and deep longitudinal groove. So much of the articular surface as is preserved agrees in structure and form with that in Plioscairus grandis ; and the extent of the angular projection behind the articular cavity to the same. The fore part of the symphysis, including the first three pairs of teeth, restored in PI. 19, fig. 4, has been subject to such violent horizontal force as to be crushed in that direction, and broken across both the upper and the under surfaces of the rest of the mandible, without having been detached from the intervening structure or tissue of the bone. The bottoms of the sockets only of the included teeth are preserved, with parts of the partitions which, here, are only from 2 to 3 lines thick. These sockets increase in size to the third. The diameter of the outlet of the fifth socket, which is the first entire one, measures 1 inch 9 lines across ; it is rather less longitudinally. The outlets of most of the alveoli are subcircular, with a tendency to a subquadrate section, with intervals not exceeding 2 lines, and they retain a uniformity of size to within four or five sockets at the end of the series, which progressively decrease in size. The total number of teeth, as shown by sockets, in each mandibular ramus, is fourteen ; of which ten occupy the symphysial part of the jaw. The upper surface of the symphysis between the first six teeth is flush with the alveolar outlets, is smooth, and slightly convex transversely. Beyond the sixth pair of teeth the intervening surface rises above the inner borders of the alveoli as high as half an inch between the ninth — eleventh pairs of sockets ; the upper surface of the hinder part of the symphysis becomes slightly convex transversely, and the pointed anterior ends of the splenials (si) enter into its composition. No part of the upper jaws of this skull of PUosaurus troclianterius has been preserved; but the quarrymen extracted the hind part of the cranium (PL 19, fig. 6). It shows a hemispheroid condyle (1) 2 inches 8 lines in basal diameter. The foramen magnum is a full transverse ellipse, 1 inch 3 lines across. The broad and low occipital surface includes the thick horizontal backwardly projecting paroccipital ridges (4), below which extend still more backward and somewhat downward the short and broad tympanies, terminated each by a condyle convex in its outer two thirds, concave transversely at the inner third : the breadth of this condyle (28) is 5 inches. The upper transverse ridge of the occiput is broken away. The parietal region (7) is formed by a lofty median vertical wall of bone, slightly expanding below to form the side walls of a miserably small cerebral cavity. I have neither respect nor inclination for undue multiplication of genera ; but the degree of difference in the number of mandibular teeth and extent of the 160 BRITISH FOSSIL REPTILES. symphysis tempts to a view of the present evidence of Pliosauriis troclianterius as testifying to something more than specific distinction from the Pliosaurus grandis. This species retains more similarity with the type Sauropterygians {PL doUclwdeirus, e.g.) in the proportions of the symphysis and of the number of symphysially located teeth. Nevertheless, modifications in these particulars are presented, though in a minor degree, by species of true Plesiosauri [compare PI. 3, fig. 2 {Plesiosaiirus dolichoderius) , with PI. 16, fig. 2 {Plesiosaurus Eawldnsii)']. A specimen subsequently submitted to me, includes the part of the maxillary bones, with eight or nine pairs of alveoli at or very near to the hind end of the series, of a smaller individual of the Pliosaurus troclianterius than that to which the lower jaw, PI. 19, fig. 4, belongs. This fragment measures 11 inches in length and 6 inches in greatest breadth. It is from the same locality and formation as the larger skull, viz. the Kimmeridge Clay of Kimmeridge. Both specimens have been liberally presented to the British Museum by the discoverer, J. C. Mausel-Pleydell, Esq., F.Gr.S. Teeth. — Of the teeth showing the generic modification above defined (p. 152) the best specimen is the subject of Plate 33. It was found in a deposit of Kimmeridge Clay, near Oxford, where fragmentary specimens had been obtained similarly testifying to the bulk and power of some old tyrant of the later Oolitic seas. In one of these specimens, the subject of PI. 19, fig. 7, the circumference of the base of the crown measures 7 inches 6 lines, equalling that of a full-sized tooth of a cachalot- whale {Phijseter macrocephatus) . Of the enamelled crown three inches are preserved and about as much of the cement-covered base, the longest diameter of which is 2 inches 2 lines ; that of the fractured end of the crown is 1 inch 3 lines. The length of an entire tooth is shown in Plate 33 ; the fractured part of the base exposes a pulp-cavity (fig. 2, c) of about 2 inches diameter, with a hard dentinal wall of from 4 to 6 lines in thickness, which gradually decreases to the broken or implanted basal ends of the tooth. The fractured part of the ci-own (fig. 2) exposes a solid and compact mass of dentine. The generic characters of the tooth are boldly pronounced. The smooth facet, defined by strong marginal ridges (fig. 3), shows a few narrower ones, encroaching upon the basal part, where the enamel gives way to the coat of cement covering the long and strong root of the tooth. This is about thrice the length of the crown, and testifies to the strength of implantation. The two sides of the crown (figs. 1 and 2) are continued into each other by an uninterrupted curve ; the ridged enamel angles between these facets and the outer smoother convex side of the tooth are well defined. The terms "convex " and " concave" refer to the longitudinal direction of the tooth ; all the sides are convex across, the outer one being the least so. The unridged surface of the enamel is finely wrinkled by KIMMERIDGIAN PLIOSAURS. 161 short wavy risings, frequently joining or reticulate, rather affecting the longitudinal course. The same character is presented by the enamel covering the contiguous parts of the other sides of the tooth, and extends furthest in that represented in fig. 1, Plate 33. The enamel, which is a mere film at the base of the crown, gains thickness towards the apex; its adhesion to the dentine is helped by numerous fine, wavy, longitudinal, sub-equidistant, linear risings on the surface of that substance. The slight outer convexity is uniform ; the inner curve is wavy, passing from the slight concavity at the crown to a slight convexity at the junction of crown and fang, then again becoming slightly concave. The subject of Plate 33, now in the British Museum of Natural History, was presented to the Trustees by the Hon. Robert Marsham, F.G.S. This tooth exemplifies its complete state of formation, the entire fang has been develojDed, and the unworn crown shows that the time had not arrived for the absorption of the root through pressure of a successional tooth, which undermining process is usually concomitant with the loss of efficiency of the dental instrument through the wear of the crown. It accordingly presents a total length of one foot, a third of which is formed by the crown, the rest by the root. This cement-covered part expands for the coronal half of its extent to a diameter of 3 inches, which is the thickest part of the tooth ; it then gradually contracts to the thin borders of the base of the pulp-cavity, where probably an additional inch of the tooth has been broken away. In excuse for the foregoing details, which may be deemed tedious, I plead the rarity of a specimen so complete as to yield the collective dental characters, and the frequency of fragments, of which any one showing part of the enamelled surface of the tooth may now have the nature of its markings and its position in the tooth-crown recognised ; thus it may throw light on the mesozoic bed, which is characterised by the present Sauropterygian genus. Finally, I may refer to the evidence of provision, in Pliosaurus, for a worn or broken lethal weapon, by a successional tooth ; and this provision, common to Reptilia, relates to the lengthened period during which the individual depends upon its teeth for sustenance. In size, in general shape, in vmiformity of character throughout the mandi- bular series of teeth, the existing Cachalots most resemble the Pliosaurs. Those whales are also deemed to be naturally long lived, but their term of life must depend on their enjoyment of the organs by which they get their food, and there is no evidence of the first fully-developed series of teeth ever being replaced by a second set. The cachalots were not favoured with the mode of supply for lengthened individual life which is exemplified in modern as in the ancient aquatic Saurians. For how many years the individual Pliosaur dominated its latitudes and preyed upon marine contemporaries we may not Ije able to determine, but its dental 1^ A. 162 BRITISH FOSSIL REPTILES. character indicates that its life period far exceeded that of the toothed whale of similar bulk. For how many centuries — generation after generation — the great Pliosaurs carried on their predatory wars some feeble conception is afforded by geological inferences from the gradual formation and accumulation of the ocean-beds which have received the dead bodies, and have now revealed to us the insoluble a:id petrifiable parts of the carcases of those cold-blooded carnivores. Sterno-coraco-scapular frame {Sauropterygia, Plate 20). — This characteristic part of the skeleton in Sauropterygia being incompletely preserved in that of Plesiosmirus dolichodeirus {8auro])teryg{a, Plate 1), its description has been deferred to the present section, when the perfect condition of the part, subse- quently worked out in another example of the same species (ib., fig. 1), serves to exemplify the modifications of the same jDart in the genus PUosmirus. In both o-enera of their order the place and function of a sternum are mainly fulfilled by the pair of coracoids (52, 52) which meet by a longitudinally extended suture (s, s) below the thoracic part of the abdominal cavity. Behind this mesial suture the coracoids diverge and terminate freely by a broad margin, each with an angle inclining laterad (52', 52') . Anteriorly the sutural portions slightly diverge, and expose the end (59') of the mesial plate representing an " episternum ;" laterally each coracoid contracts in length, becomes thickened, and presents two roughened articular siirfaces ; the hinder one (}') contributes the corresponding portion of the articular cavity for the humerus, the fore one (ch) joins the scapula (51) by the suture laterad of which the scapula contributes the free portion {h) of the glenoid cavity. In Plesiosmirus (fig. 1) the hinder end of the scapula, which is the thickest part of the bone, is thus divided pretty equally between its coracoidal (ch) and humeral (/') articular surfaces; both are rough or " syndesmosal," the latter least so. In advance of the surfaces {h, ch) the scapula thins and contracts chiefly by a strong margino-mesial concavity contributing the outer border of the " coraco- scapular vacuity " (c s). The outer, thicker border of the scapula, in Plesioscmrus, (fig. 1) is straight, and the bone extending forward expands to unite with the episternum (59) by the suture («A). The episternum (59) presents anteriorly a mesial notch, from each angle of which a thicker border extends outward and backward to its sutural union {s li) with the fore end of the scapula. At this union the episternum contracts and is continued backward to join the coracoids, passing a short way internal to them, and appearing externally as a pointed end of the bone at the fore part of the narrow mesial interspace of the coracoids, which interspace interrupts, anteriorly, their extensive mesial sutural union, s, s. Thus the sterno-coraco-scapular frame presents an anterior and a posterior notch and a pair of subcircular vacuities. The above-defined characters of this KIMMERIDGIAN PLIOSAURS. 163 portion of the skeleton, save that of the scapular clement, are common to both the genera of the Sauropterygia. The chief and suggestive modification of the mass in the Pliosam^ian genus is the retention of a typical character of the scapula which is lost in Plesiosaurus, the production, namely, of the part of the blade-bone (PI. 20, fig. 2, si x) latcrad and dorsad, where it terminates freely. This portion represents the main body of the scapula in higher vertebrates ; but, as in the allantoic or abranchiate group,^ (Beptilia, Aves) without expanding. The portion of the scapula common to both the present extinct genera, which contributes its share (h) to the glenoid cavity, is separated, in Pliosaurus, from the free portion (51 ^r) by the notch (n). In advance of this the Pliosaurian differs from the Plcsiosaurian scapula by its greater relative breadth, extending its sutural border (sA) mesiad so as to touch or join the fore end of the coracoid (so). The coracoids retain their large proportional size, but have a less even or flattened outer surface ; mesially they bidgo to their common suture (s, s), giving more room to the ventral cavity ; and, at the transverse parallel with the borders, which they contribute to the vacuities (c s,cs), they bend dorsad, or inward, suddenly contract (sc) ; but contribute, as in Plesiosaurus, the mesial border of those vacuities, and articulate, underlapping it, with the hinder end of the episternum (59'). In thus determining the homologies of the constituents of the complex bony buckler in 8auropteri/gia, I have exhausted every subject of comparison at my command in other Reptilian forms, both fossil and recent. The degree in which the abdominal surface is defended by bone in Sauroptcrygia resembles that in Ghelonia. But, as I have elsewhere^ shown, by dermal ossifications chiefly, these not answering to the endoskeletal elements which have been modified to that end in the subjects of the present section. Pliosaurus pobtlandicus, Owen. Smiropterygia, PI. 19, fig. 8. The generic distinction from Plesiosaurus, indicated by the term Pliosaurus, and originally suggested by characters of the teeth and the cervical vertebrse, is further confirmed by the structure of the bony framework of the paddles. The modifica- tion in question, like the fore-and-aft compression or shortness of the cervical centrums, exemplifies the nearer resemblance, I will not say affinity, of Pliosaurus to Ichthyosaurus ; the segment of the natatory limb which answers to the anti- 1 ' Anatomy of Vertebrates,' vol. i, 1S66, pp. 6, 7. 2 'Philosophical Transactions' for 1849, "On the Development and Homologies of the Carapace and Plastron of the Chelonia. 164 BRITISH FOSSIL REPTILES. brachium y) plates; they increase in breadth after the sixth; the first bears the impression of the triradiate line which mai'ks the division between the first (m i) and second {m 2) marginal scutes, and the first (« i) vertebral scute. There is no nuchal scute. The second, third, and fourth marginal plates are slightly overlapped by the first costal scute (c l). The antero-posterior breadth, in comparison with the transverse breadth, is greater in the costal scutes of the Pleurosternon ovatum than in those of the Pleurosternon emarginatum. The number of marginal scutes is twenty-four, twelve on each side (m 1 to m 12). The fore part of the plastron appears to have projected in advance of the carapace, as is indicated by the plate of bone marked (e s) in Plate 57. The length of the carapace of the specimen of Pleurosternon oratmn here described is 19 inches 6 lines; its breadth is 14 inches 6 lines. It is very slightly convex, with the margins a little raised. The feeble sculpturing of the outer surface of the carapace resembles in general character that of the other species of Pleurosternon. Pleueosternon LATrscuTATDii, Oicen. Plate 58. The species represented by the specimen of mutilated carapace, here figured, differs from all the other recognised species of the genus by its distinct nuchal scute {ch), by the small relative size of the first vertebral scute {v l), and by the 24, 186 BRITISH FOSSIL REPTILES. great relative size, more especially the superior breadth, of the three succeeding vertebral scutes (y 2, v 3, v 4). The boundary lines, indicating the forms and disposition of the horny scutes, are proportionally larger and deeper than in the other species of Plcuro>ifernon which have come under my observation. The sutures uniting together the different elements of the carapace are more dentated or wavy, more especially the suture uniting the nuchal plate with the first neural plate and first pair of costal j^lates. The neural plates, from the first to the seventh inclusive, are similar in form, six-sided, with the antero-lateral sides the shortest; the eighth neural plate is the smallest, is four-sided, and broadest behind ; the ninth and tenth neural plates are remarkable for their great breadth . The transverse extent or length of the costal plates is considerable, in accordance with the great breadth of the carapace : the eighth costal plate, in this respect, differs considerably from its homologue in the other species of Pleuro- sternon. The second marginal scute is not produced backwards between the first vertebral and first costal scute, but, like the first and third marginal scutes, has its antero-posterior diameter much less than the diameter in the direction of the periphery of the carapace. The first (c 1) and fourth {e 4) costal scutes differ considerably in their forms and proportions from those in Plates 53, 55, and 57. The outer surface of the osseous parts of the carapace of Pleurosternon latiscutatum is minutely punctated and rugose, except near the sutural borders of the several pieces, where it is impressed by rather coarse parallel striae, directed at right angles to those borders. Genus — Platemts. Platemts Mantelli, Owen. Plate 52, fig. 1. Amongst the Chelonian Fossils from the "Wealden strata of the Tilgate Forest, in Sussex, are certain specimens which resemble the flat species of Emydian, or terrapene, discovered by M. Hugi, in the Jura limestone at Soleure. Both the Jura species and the Wealden Chelonites in question are referable to the ' pleuro- deral' section of the great tribe Faludinosa, as arranged by Messrs. Dumeril and Bibron ;' and, in that section, to the genus Platemys. The most intelligible fi'agment in the British Museum, is that element of the plastron — the hyosternal, which is figured in the above Plate. The proportions of this bone indicate that the plastron of the Platemys Mantelli consisted of the » ' Erpetologie,' 8vo, 1S35, torn, ii, pp. 17L', 372. WEALDEN CHELONIANS. 187 ordinary nine pieces : where the accessory pair of mesosternal pieces is introduced, both the hyo- and hypo-sternals have relatively less antero-posterior extent than the fossil in question shows. Platemys, sp. dub. Plate 52, fig. 2. A second species of Wealden Platemys is apparently characterised by a some- what broader plastron, and by a greater relative thickness of the bones composing both this and the carapace. Without the latter difference, the proportionally broader plastron might be merely the sexual distinction of the female of the first species. Some difference, in the shape of the axillary notch of the hyosternal further induces me to regard the fragmentary Chelonites in question, of which a hyosternal is figured in Plate 52, as belonging to a second species of Wealden Platemys. Platemys Dixoni, Owen. Plate 52, fig. 3. A Platemydian specifically distinct fi'om either of the above is more unequi- vocally exemplified by the sternal element represented in figure 3 ; the matrix having been carefully removed from the outer surface of this fossil, the linear impressions which have divided the humeral from the pectoral scute, and this from the abdominal scute, are clearly shown. The positions of these transverse grooves accord with those in the hyosternal of the Bmydians, having the usual number (nine) of plastral elements : and the hyosternal character of the fossil is further shown by the oblique border cutting off the inner angle of the anterior end, for articulation with the entosternal element, (This end has been figured downwards in the plate.) The axillary groove is narrower than in the above- described species ; and the whole bone seems to have been longer in proportion to its breadth. It is from the Wealden of Tilgate Forest, and formed part of the Collection of the Author of the instructive Work, ' On the Cretaceous and Tertiary Formations of Sussex,' Frederic Dixon, Esq., F.G.S. Genus — Chelone. Chelone costata, Oicen. Plate 51. From the Wealden Clays of Tilgate Forest have been obtained many frag- mentary Chelonites, indicative of species representing two of the actual families of the order, viz. Paludtnosa and Marina ; and such, therefore, as might be expected ]88 BRITISH FOSSIL REPTILES. to be met witli in the deposits of a large estuary. I commence tlie description of these Wealden Chelonites by those which indicate a species of the marine family. Portions of the carapace and plastron, and bones of the extremities of a large species of Turtle, some of them indicating individuals with a carapace nearly three feet in length, form part of the Mantellian collection, purchased by the Trustees of the British Museum, and now in the Museum of Natural History, Cromwell Road. After comparison of these specimens I have come to the conclusion that the Wealden species differ from Chelone imbricata, Chelone carinata, and other recent species, in as great a degree as do most of the Eocene Ghelones, in the greater extent of ossification of the costal interspaces and of the plastron. A characteristic portion of the great Wealden Turtle is represented, of the natural size, in Plate 51. It includes the second and third marginal plates, and considerable portions of the first and second costal plates, with the connate portions of the pleurapophyses, or vertebral ribs. These are remarkable for their breadth and prominence, and have suggested the name proposed for the present species. In the same plate are represented a mutilated right iliac bone (fig. 3) and the right femur (fig. 2) of, probably, the same species of Turtle. These, also, are from the Wealden formations of Tilgate Forest. Figure 4, Plate 52, gives a view of the inner surface of the left hyposternal, half the natural size of, probably, the same species of Chelone. It is embedded in a slab of Wealden stone. As compared with existing Turtles, the ossification of the plastron is more, advanced or more extensive, the rays of bone from the outer and inner free borders of the hyposternal being shorter and their interspaces more filled up. A nearer approach is thus made in this Wealden species, as in some of the Eocene Turtles, to what may be regarded as the more general type of the Chelonian carapace. Chelone gigas, Owen. Plates XXX, XXXI. In the course of the determination and description of the fossil remains referable to the marine genus Chelone (Turtles) fragmentary specimens from the Eocene clay of the Isle of Sheppey indicated the then existence of a species, as shown in fig. 5, Plate 40, much larger than those described in Vol. I, pp. 7 — 44. Recently, however, have been acquired from that locality, for the Palajontological Department of the British Museum of Natural History, remains of this giant of the Family Marina, of which I have selected for illustration the skull, represented in Chelonia, Plate XXX, of the natural size, viewed from above ; other views of TRIASSIC LA.BYRINTHODONS. 189 the same remarkable fossil being given, reduced to one tliird the natural size, in Plate XXXI, Vol. IV. By a comparison o£ these figures with those of the skulls of other extinct species which have left their remains in the same formation and locality it will be seen that besides difference of size, which is sufficiently remarkable, there are also proportional characteristics which compel a reference of the giant of the marine family to a distinct species. The entire skull is more flattened and depressed. This is shown in figs. 1 and 3 of PI. XXXI. If they be compared with figs. 1 and 3, PL XI (Ghelone cuneicejys). Vol. II, which offers the nearest approach to Ch. gigas in this character, the difference will be obvious. It is not wholly due to posthumous pressure, although this has produced a partial dislocation, as shown by the slightly bent super-occipital in fig. 3, and the shape of the orbit in fig. 1, PI. XXXI. The plane of the nostril is wholly upon the upper surface of the skull, and is relatively nearer to the orbits, and further from the anterior premaxillary part. A transverse line across the back part of the nostril crosses the orbit very little in advance of the hind part of the anterior third of that cavity, a rare relative position of the nostril which could not have been induced by mere pressure. The orbits open upon the anterior half of the skull. The premaxillaries are produced beyond the nostril for a greater relative extent than in any of the extinct kinds, in some of which, for example Ghelone longiceps (PL XII, fig. 2), Vol. II, the plane of the aperture is the same with that of the fore slope of the skull. In the recent Turtles, Ghelone mi/das, for example, the nostril is terminal, and its plane almost vertical, and it opens wholly in advance of the orbits. These apertures are relatively smaller in Ghelone gigas than in any recent and in most extinct species. All the cranial characters of the marine family of the order Ghelonia are present in the gigantic extinct species. Order— LABYRINTHODONTIA. Genus — Labyrinthodon, Oiven. A knowledge of the chief character of the present Oi'der and Genus was derived from examination of portions of petrified teeth found in a quarry of the New-red Sandstone at Coton End and Guy's Cliff, Warwickshire, and transmitted to me for determination and description by Murchison and Strickland.^ The specimens received indicated a tooth of the common canine character, but straight 1 See their paper iu the ' Transactions of tlie Geological Society,' 4to, vol. v, Part ii, 1840. 190 BRITISH FOSSIL REPTILES. Fig. 1. — Section of tooth of Lahyrinlhodon. and with a subcircular transverse section : the surface was traversed by close-set longitudinal lines, seemingly indicative of fissures. Being at tbat tinie engaged in a study of the teeth of Vertebrates, I submitted the fossils to microsco])ical scrutiny, and great was my surprise to see, in a transverse section, the structure figured in the subjoined Cut, fig. 1. The question which chiefly interested my geological friends was whether the sandstone was, as tbey suspected, of Triassic age, and might be eqiiivalent to the German " Keuper Sandstone." Now, from this sandstone had been obtained fossils of a large Vertebrate species, referred by Professor Jaeger^ to a genus he termed Mastodon-saurus. Of this species one of the fossil teeth presented the same conical, transversely circular, shape and longitudinal striation of tbe Warwickshire fossil. I therefore wrote to the author requesting the favour of a tooth of the Mastodonsaurus, which was promptly and kindly granted. Of this tooth slide-sections for the microscope were prepared and a labyrinthie interblending of the dental tissues was displayed, identified with the structure so unexpectedly brought to light in the fossil teeth from the Warwickshire Trias. Subsequent acquisitions of fossil remains, for which I am indebted to Dr. Lloyd, of Leamington, have enabled me to add the following illustrations of the osteological characters of the extinct form characterised by the labyrinthie structure of its teeth. Species — Labyrinthodon Jaegeri, Owen. The first of these fossils here described indicated a species as large as the type of Mastodonsaurus. It consisted of portions of two mandibular rami [Batrachia, Plate 2) from diff'erent individuals. One, figs. 1, la, includes the angular, articular, and hinder part of the dentary elements of the lower jaw, with portions of a score of relatively small conical, nearly equal-sized teeth. The angle of the jaw termi- nates obtusely, and is produced about three inches behind the articular surface. ^ ' TJeber die fossilen Reptilien, welche iu Wiirtemberg aufgefuadeu wordea sind,' 4to, 1828, pp. 35, .38, tab. 4 and 5. TRIASSIC LABYRINTHODONS. 191 From this the bone gradually decreases in vertical extent to its broken fore end. In the second and larger fossil (fig. 2) the exterior of the lower three fourths of the bone is strongly sculptured by obtuse interrujDted ridges, mainly radiating from the lower border below the articular surface (fig. 2). The inner side of the ramus (fig. 2) is comparatively smooth, and shows the termination of the depression which lodged the liiud end of the dentary element. Portions of the maxillary bone and teeth of this species are shown in figs. 4, 5, and 6, Plate 4. Species — Lahijrinthodon leptognatlms. (Batrachia, Plate 3.) Of the fossils referred to the above species the most instructive was the portion of skull represented in figures 1 and 2, of the natural size. Careful removal of the stony matrix exposed, on the palatal surface (fig. 2) a broad divided vomer (h), contributing a somewhat larger proportion to the roof of the mouth, than the divided vomer characteristic of existing Toads and Frogs. The position and rela- tive size of the inner or palatal nostril (c) added to the batrachian characters. Anterior to this part of the palate was the base of a tooth, which, compared with the row of maxillary teeth, might be termed a tusk. The labyrinthic structure was instructively shown in a section of this tooth. So much of the upper wall of the skull is preserved as to show the broad, flattened shape of its facial portion ; but the extent of the maxillary and nasal bones composing the roof presents a marked distinction from the framework of the similarly shaped skidl in existing broad and flat-headed Batrachians. In these the maxillaries have the form of elongate styles, attached by a slightly expanded fore end, and terminating behind in a free point : they are, also, edentulous. The outer surface of the broad facial part of the skull of Lahyrinihodon is sculptured in a degree recalling that of the outer surface of the lower jaw of the huge species above described. Irregular grooves and sinuses are divided by corresponding risings. An angular furrow runs nearly parallel with the alveolar process, a little above it, defining it from the broad upper flat surface of the skull ; a second less angular furrow inclines one side as it extends forward. The alveolar part of the fossil includes thirty-one sockets, the foremost, lodging the base of a tooth three times the size of the next, which commences the series of smaller teeth, gradually decreasing in size as they extend backwards. A side view of the above- described fossil is given in fig. 3, in which a indicates the tusk of the outer dental series and h that of the vomer. The dental character of the present species is more fully shown in the con- siderable proportion of the left mandibular ramus, figured in Batmchia, Plate 4, 192 BRITISH FOSSIL REPTILES. figs. 8 and 9. The conformity of this cliaracter with that shown in the preserved portion of the upper jaw will be appreciated by the side-view of that portion introduced in fig. 7 of Plate 4. In the vacant mandibular sockets, corresponding to the upper teeth in place, are germs of successioual teeth, more or less advanced in formation. The number of sockets in the single alveolar series is not less than fifty : to which add a much larger canine or tusk at the inflected symphysial end. A Batrachian mandibular character is exaggerated in Labyrinthodon by the extension of the angular element of the jaw along the under part of the ramus to the short symphysis. The ' harmonia,' or toothless suture indenting the outer surface, indicates the proportions of the angular and dentary elements contributed to that surface. Figures of the best preserved maxillary and mandibular seinal teeth, slightly magnified, are given in advance of the views, natural size, of the upper and lower jaws of Labyrinthodon leptognathus. Labyrinthodon pachygyiathus, Owen. (Batrachla, Plates 3 and 4-.) The portions of upper jaw on which this species is founded are represented of the natural size in figures 4, 5, 9, and 10 of Plate 3. The outer surface of the portion preserved of the maxillary shows the characteristic coarse sculpturing ; part of the vomerine nostril is indicated at c, figs. 9 and 10. The alveolar and part of the palatal processes of the maxillary afford the subjects of figs. 4 and 5 ; and the crown of the best preserved maxillary tooth is represented in fig. 6. Figs. 7 and 8 are portions of a mandible, but the characters of this bone and of its teeth are exemplified in figs. 1, 2, 3 of Batrachla, Plate 4, from parts of the right ramus. The outer surface of the dentary is traversed by a longitudinal groove midway between the upper and lower borders, indicative of the proportions of the dentary and angular elements thereto contributed. The part of the outer surface of the angular in the hinder portion of the mandible of Lab. pachygnathus is broken away. On the inner surface of the fore part of the ramus (fig. 2) is shown the pointed termination of the splenial element, which extends to near the symphysis ; an upper view of the same fore end of the ramus is given in fig. 3. The mandibular dentition is instructively shown in the present specimens. The small serial teeth exhibited more or less entire, or indicated by sockets, in the two portions of jaw, are not fewer than forty. The symphysial end of the ramus supports two much larger tusk-like teeth, with indication of a third of less size, but exceeding that of the serial teeth. Of these the crown is best preserved in the posterior portion of the ramus (fig. 1 ) , which had been detached from the rest. The TRIASSIC LABYRINTHODONS. 193 vacant sockets between the teeth in place show more or less advanced beginnings oE successional teeth. The figures of the mandible and teeth are of the natural size. Vertebrce of Labyrinthodon. (Batrachia, Plate 5.) The proportions of a vertebral fragment, associated with the above-described mandible, in the Trias of Coton End, lead me to refer it to the species pachy- gnathus. Fig. 1, showing a portion of the fore articular surf ace of the centrum , with the coalesced base of the neural arch, gives the moderately concave character of that surface. The upper view of the same vertebral fragment (fig. 2) determines the fore and hind ends by the bases of the zygapophyses of the coalesced neural arch fortunately remaining. The base of a broad, depressed diapophysis is also shown, and is further exemplified in the side view (fig. 4). The hind articular surface of the centrum has suffered fracture, forbidding determina- tion of its natural shape. The minutely cellular, almost compact, texture of the bone is displayed by the fractured surface in fig. 3. A better preserved vertebra, in size referable to Labtjrinthodon leptognatlms, affords the subjects of figs, 5—8. The degree of concavity of the fore surface is shown in fig. 7, in which the centrum is associated with so much of the neural arch as exhibits the position and shape of the prezygapophyses which received the articular ends of the postzygapophyses, unfortunately mutilated, with the corre- sponding articular surface of the centrum, as shown in fig. 8, In the characters of the vertebra from a part of the trunk, as exemplified in the specimens from two of the British species of the present singular genus, we find the Labyrinthodon superadding modifications to the vertebrae of the highest existing Batrachia (toads, frogs, salamanders), which, as in the dental and osteological characters next to be noticed, manifest an association of Reptilian (Crocodilian, Dinosaurian) features with an essentially Batrachian organisation. The portions of ribs which have been recovered show these bones to have been of greater relative length and curvature than in any existing Batrachians, and in the character of size they accord with that of the articular process and surface developed from the neural arch. The subject of figs. 9 and 10, in Plate 5, might well have been interpreted, if found alone, as evidence of an extinct Reptile of higher grade than a salamander, — to an Icthyosaur, for example. It is plainly a sternal bone, showing articidar concavities for a pair of clavicles ; or it may answer to that part of the complex scapular arch in Reptiles which has been named " episternum." The locality of the fossil and its associations with unquestionable remains of Labyrinthodont reptiles support its reference to that genus ; and, from its size, to the species leptognatlms. The stem or body of the bone thins off as it recedes fi'om the 25 194 BRITISH FOSSIL REPTILES. articular process to a flat plate, from whicli tlie end is broken away. The advanced tliicker end expands and extends into cross pieces, at right angles, each with an articular depression indicative of clavicles. Now, these bones, which are absent in Crocodilia, are present in higher Batrachia, and, in Bitfonidce, their mesial extremities rest upon the expanded fore end of an episternal bone; it is not, however, curved lengthwise as shown in fig. 10, in Labyrinthodon, a curvature which indicates a greater vertical capacity of the fore part of the thoracic- abdominal cavity. Sumerus. The fossil from the Trias at Coton End, of which four figures (11 — 14) are given in Plate 6, is the proximal portion of a humerus. The moderately-convex, proximal, articular end (fig. 14), from which extends the beginning of a well- developed deltoid ridge, and the characters of the shaft shown by the surfaces divided by that ridge, are more like those of the humerus of a toad than in that of any Lacertian, Chelonian, or Crocodilian Reptile. The bone had a medullary cavity of the width shown in fig. 13. If this limb-bone should belong to the same species as the ilium (figs. 16 and 17) the disproportion of size in the fore and hind limbs would be as in the anourous Batrachians ; but I have received evidences of the tail of Labyrinthodon. Great part of the ilium is devoted to the formation of a large acetabular cavity ; this is of an oblong form, extending in the long axis of the bone; its margin, elsewhere sharp, is smoothed away at the base of the iliac body, which becomes narrow and compressed as it recedes. The chief distinctive character is the process above the acetabulum, from which it is separated by a smooth concavity ; this process is compressed as it rises, and is bent forward, ending in an obtuse point. A process of a different shape rises in a similar position above the acetabulum in the frog. From the superacetabular process the ilium is continued forward, and terminates in a thick subtruncate surface a few Unas in advance of the acetabulum. The extent to which this ilium is articulated to the vertebra, at least three in number, which may be regarded as sacral, is shown in the mesial view of the bone given in fig. 17; the superacetabular process and the hinder slender production contributing to the vertically concave articular surface. Of a femur, corresponding in size, in any degree, with the above ilium, I have, as yet, received only the hemispherical head, represented in fig. 18. But in a group of bones of a small, or possibly young. Reptile from the New Red Sandstone (Trias) of Lymington, with the distal articular end of a femur (Batrachia, Plate 0, fig. 1,/), were associated a tibia (t) and a humerus (''), plainly indicating a great disproportion in size between the fore and hind limbs. TRIASSIC LABYRINTHODONS. 195, I subsequently found that a fossil from Warwickshire Trias, figured in Plate XXVIII, fig. 9, of the above-cited paper by Murchison and Strickland, was a terminal phalanx showing a Batrachian character in the absence of the usual modification for the insertion or attachment of a claw. Labyi'inthodon (Anisopus) scutulatus, Owen. Returning to the group of bones in Plate 6, figs. 1 — 5, I found them to belong to a small reptile with the biconcave system of vertebrae, but which, from the length, structure, and form of the long bones of the extremities, must have been of terrestrial rather than marine habits, and which had the skin defended by numerous small rhomboidal bony scutes, with a smooth central surface, and with the outer surface sculptured by three or four longitudinal ridges (fig. 6). This reptile had the hind legs twice as long and as strong as the fore. The humerus (fig. l,k) is convex at the proximal extremity, it is expanded both at this and the distal extremities, and is contracted in the middle. There is a portion of a some- what shorter and flatter bone, bent at a subacute angle with the distal extremity of the humerus, and which presents the nearest resemblance to the anchylosed radius and ulna of the frog. The proximal extremity is wanting in the femur (fig. 1,/), the remnant of the shaft is slightly bent, and is subtrihedral ; its walls are thin and compact, and include a large medullary cavity. Both tibise exhibit that remarkable compression of the distal portion of the shaft which characterises the corresponding bone in the anourous Batrachia, and both likewise exhibit the longitudinal impression along the middle of the flattened surface. The vertebrae (figs. 2, 4, and 3 magnified) are biconcave, with these surfaces sloping obliquely from the axis of the body, as in the dorsals of a frog, indicative of habitual curvature of the part of the spine formed by them. The aquatic Salamanders, including the gigantic species from Japan, have both ends of the vertebral body concave, but more conical than hemispherical, as in the present fossil, which in this respect resembles the Labyrinthodont vertebrse (figs. 1 and 7) in Plate 5. Portions of ribs associated with the above-described fossils showed them to be longer and more curved than in t]>e existing remnants of the Batrachian type. The Leamington fossil also exhibits a character, in the small, bony dermal sculptured plates, not yet found in the Warwick or Wirtemberg Labyrinthodons, which seems to remove it from all Batrachia— the naked reptiles, as they are emphatically termed — and to approximate it to the Loricated Order. These scutes (fig. 5) form a suggestive instance of the Crocodilian affinities of the Leamington Batrachian ; we have already seen the same affinities mani- 196 BRITISH FOSSIL REPTILES. fested in other parts of their organisation by the larger Labyrinthodons. As these detached superficial bones are the most liable to be separated from the fragmentary skeleton of the individual they once clothed, the mere negative fact of their absence, when so small a porportion of the bones of the trunk of any Labyrinthodon has yet been found, is insufficient to prove a difference of dermal structure between the Leamington and Warwickshire species. No anatomist, indeed, can contemplate the extensive development and bold sculpturing of the dermal surface of the cranial bones in the Labyrinthodontes pachygnathus and leptogiiatlms, without a suspicion that the same character may have been manifested in bony plates of the skin in other parts of the body. And granting that this structure existed, to what extent, it may be asked, does it affect the claims of the Labyrinthodon to be admitted into the order of Batrachia, in which every known species is covered with a soft, lubricous, and naked integu- ment ? To this question it may be replied, that the skin is the seat of the most variable characters in all animals ; and, if considered apart from the modifica- tions of the osseous and dental systems, is apt to mislead the naturalist who is in quest of the real affinities of a species. Suppose, for example, that the existing Chelonian Reptiles were exclusively mud-tortoises, or with a soft and naked skin, as in the species of Trionyx and SijJiargis, would the discovery of the osseous carapace of a true Testudo, in a fossil state, in connection with a skeleton in other respects essentially corresponding with the modifications exhibited by a Trionyx prohibit the association of the fossil in the same order of Reptiles with the Trionyx, because of the indication of the scutes ? It unquestionably ought not to affect such a determination. And so with respect to the Labyrinthodont Batrachia ; if all the species have pushed their affinities to the Crocodilians so far as to have had their trunk defended by bony dermal plates, yet their double occipital condyle, their comparatively simple lower jaw, their large vomerine boiies and teeth are decisive of their Batrachian nature. In the " Alaunschiefer of the German Keuper " was found the occipital part of a fossil skull, with a dovible condyle to which the name Salamandroides giganteus was given by Jaeger. I am of opinion that, with the Mastodonsaurus, it was also a Labyrinthodont. These extinct forms deviated from existing Salamanders in the crocodilian development and sculpturing of the cranial bones, and iu having dermal osseous plates. Finally, I have to offer remarks, on the Batrachian affinity indicated by their foot-prints. Since the above-described fossils were submitted to my examination impres- sions and reliefs of impressions of foot-prints have been found on slabs of the New Red Sandstone in different British localities, proclaiming the primitive TRIASSIC LABYRINTHODONS. 197 Fig. 2. Foot-priuts of Labyritithudon. k ■li^. i"^ S^ ■i'-'/, ''11,1'' \ \ plastic condition of such stones wlien so impressed at low water and receiving successive tidal deposits of the same fine sand. Impressions and reliefs of such prints have been traced for many steps in succession, in one instance of which a portion is represented in the adjoining Cut. They have been noted in Triassic formations of Warwick- shire and Cheshire, and in a quarry of whitish quartzose sandstone at Storeton Hill, a few miles fi'om Liverpool. Some are hollow, as they were impressed, others are in reHef, being natural casts ; always, respectively, on opposite surfaces of the sandstone slabs. Such impressions or " ichnites " indicate vagrants of different sizes. Those left by the hind foot, in the largest kind, are eight inches in length, five inches in width ; and near each, at a regular distance — about an inch and a half in advance — is a smaller print of the fore foot, four inches long and three inches wide. The footsteps follow each other in pairs at intervals of about fourteen inches from pair to pair. The large (hind) as well as the small (fore) steps show the thumb-like outer toe alternately on the right and left side, each step leaving a print of five toes, in which there are no indication of claws. Foot-prints of this kind were first observed in Saxony, near Hillburghausen, in quarries of a Liassic sandstone. Dr. Kaup, who (1836) described them, gave the name of Gheirothermm to the animal that made them, in reference to their resemblance to the impression left by a human hand. But, led by a like disproportion between the fore and hind limbs in the kangaroo, he conjectured that they might indicate an extinct form of the Marsupial order of quadrupeds. In Bklelphys, how- ever, the thumb is on the inner, not the outer, side of the hind foot, and is on a line with the other toes in the fore foot. Decisive evidence of a species of Mammal being in existence at the Triassic epoch has since been had ; but the remains of Tritijlodon ^ have not yet revealed the structure of the feet. ' 'Quarterly Journal of the Geological Society,' February, 1884, p. 146, PI. VI. ^ V¥i ar.^^T^^l k 'y*^*- '?i. V s i; } 198 BRITISH FOSSIL REPTILES. The Cheirotlierian Ichnites resemble the foot-prints of a Salamander in having the outer toe of the hind foot projecting at a right angle to the line of the mid- toe; they recall the foot-prints of the toad in their unequal size. The fossil remains, above described, from the Triassic deposits and localities exhibiting the Cheirotherian impressions, justify the conclusion that they have been made by a cold-blooded rather than a mammalian Marsupial animal, and by a species of the class which includes Batrachians with a similar disproportion between the hind and the fore limbs. On this hypothesis it is not less evident that the impressing vagrant was quite peculiar and distinct from any known Batrachian or other reptile in the form of its feet. The analogy of the Crocodilian reptiles would indicate the short and freely-projecting digit to be the outer or fifth toe, whilst the closer corre- spondence of the Batrachian feet would prove it to be the inner or first toe ; but the thickness, relative size, and position of the remaining toes are peculiarities of the Cheirotherian footsteps. Thus, in LabyrintJiodon we have a Batrachian reptile, and one that differs very remarkably from all known Batrachia and every other reptile in the structure of its teeth : it is also a Batrachian, which, with strong aflinities to the Sauria, appears to have presented the same inequality of size between the fore and hind extremities as does the so-called Cheirothere : and both the footsteps and the fossils are peculiar to certain members of the Triassic formations. May we not then be justified, upon this evidence, in adding the name Gheirotherium to Masto- donsaurus and Phytosaunis among the synonyms of the genus Labtrinthodon ? I have already alluded to footsteps of a different but somewhat allied form, as being probably those of the Lab. leptognatlvus. These footsteps actually occur associated with those to which the name Cheirotherium has been given on the same slab, in the sandstone quarries at Storeton, but are more Crocodilian in their character. Since my first acquaintance with this type of reptile I have received and described other specific and generic forms from Hindostan, America, and, as in the instance of the species of Bi/tidostens, from a Triassic formation at the Cape of Good Hope.^ Here the progress of time compels me to conclude the present work. To those who encouraged me in the undertaking I plead the intervals elapsing between the acquisitions of the subjects, and the frequent indication of a new form of Extinct Reptile by a fragmentary fossil, calling for further research in the locality, and lapses of time before additional evidences justified a reconstruction and a name for the long lost monster. ' ' Quarterly Journal of the Greological Society,' Marcli, 188-1. CONCLUSION. 199 A glance at any summary of tbe Beptilia still maintaining an existence in Great Britain^ will impress the contrast between them and the numbers, the hugeness, the strange modifications of such cold-blooded air-breathers which lived on the lands and swam by the shores of the Mesozoic and Eocene worlds. These discoveries suggest, also, the most probable correspondence of the climate of the ancient continents at those epochs with that of the countries where the crocodile, the alligator, the ghavial, the boas, and the larger chelonians still find conditions of existence. Nevertheless, what most impresses the writer is a sense of the fragmentary nature of the present contribution to a restoration of such forms of past life, and the conviction of the extent of the field which, especially in the Mesozoic strata of our island, still remains for the cultivation of the Reptilian branch of Pala30ntology. 1 Bell's ' British Eeptiles,' 8vo., V. Voorst. THE ENP. ■-■'■■'■' r'--''::''i'^v-^-i'^--'i^^^^