considerable variations in size, has been named Leptoptilus falconeri; while a cervical vertebra from the Punjab indicates another genus of the same group. Another cervical vertebra from the same district has been provisionally referred to Mergus. The Ratitæ appear to have been represented by two species, one of which was a true ostrich (Struthio asiaticus 1), and is known by several bones of the leg and foot, and some cervical vertebræ; while the other was a three-toed form of which the genus has not been determined.

CLASS III.-REPTILIA.

I. TERTIARY.

Pleistocene. The reptilian remains from the Karnul caves comprise a tooth of a species of Crocodilus; numerous bones of the existing Madras monitor (Varanus dracana); and, in the Ophidia, vertebræ of Python molurus, of the cobra (Naia tripudians), and of a large snake provisionally identified with Plyas mucosus. Some other small ophidian, as well as chelonian, remains have not been even generically identified.

The reptiles of the older pleistocene are still very imperfectly known, but it is probable that they all belong to living Indian species. From both the Jamna and Narbada beds specifically indeterminable remains of crocodiles have been obtained. A complete specimen of the carapace of Pangshura flaviventris from the Narbada is in the Indian Museum, and serves to connect the living with the Siwalik form, and also shows that the range of the species once extended over the greater part of India; a less perfect specimen in the same Museum may also belong to this species. A portion of the plastron of a Batagur from the Narbada valley has been provisionally referred to B. dhongoka, now found in that river. A fragment of the carapace of a Trionyx, from the same deposits, probably belonged to T. gangeticus, and a cranium in the British Museum gives more certain evidence of the occurrence of that species. Pliocene and (?) upper miocene.—-The Chelonia of the Siwaliks although still imperfectly known, are represented by a considerable number of forms. In the Testudinide the gigantic Colossochelys atlas is the largest of all known forms; it is distinguished from Testudo by the non-union of the pygal plates 3 of the carapace, and by the production of the plastron anteriorly into a pair of cornua, supported on the ventral aspect by a strong triangular keel on which the gular plates are borne; the length of the restored carapace in a straight line is 8 feet 4 inches. Four other forms. of gigantic land-tortoises are indicated by remains which are not sufficiently perfect to admit of generic determination. The first of these species is about one-half larger than the living Testudo elephantina of Aldabra, and has an epiplastron intermediate in structure between that of Colossochelys atlas and that of the existing Indian Manuria emys. The second, which may be identical with the form to which the name Cautleya annuliger has been applied, is about one-fourth larger than T. elephan

'A limb-bone of this species, and a specimen which is apparently not avian at all, were made the types of the genus Megaloscelornis. Other bones were referred to Dromaus but these also turn out not to be avian.

2 Originally described as P. tectum.

The term plate is applied to the horny epidermal covering, and scute to the subjacent bony layer.

tina. The third is of nearly equal dimensions, but has a very different epiplastron ; while the fourth is considerably smaller, and appears more nearly allied to the existing land-tortoises of India and Burma. In the Emydida, Clemmys is represented by several forms, six of which have received distinct specific names. Of these C. sivalensis is allied to the existing C. crassicollis of India and the neighbouring regions, but has no nuchal plate; C. hydaspica is an allied form in which the nuchal plate is present; while a third member of the same group is found in C. theobaldi which has an unusually depressed carapace, with a first vertebral plate of very remarkable shape. C. punjabiensis is a form with a bell-shaped first vertebral plate, in which respect it resembles certain North American species of the genus; its hinder vetebrals have not the 'mushroom-shape,' characteristic of the three preceding species. A fifth unnamed species comes so close to the existing Indian C. trijuga, that it may be pretty safely regarded as the ancestor of that species, if indeed it be not identical. From Perim Island a shell with quadrangular vertebral plates has received the name of C. watsoni 1; while C. palaindica of the Siwalik hills appears closely allied to the tricarinate C. hamiltoni of India, of which it may have been the ancestor. An unnamed form from Perim Island may perhaps indicate an eighth species of the genus. The characteristic Indian genus Pangshura is represented by a form provisionally identified with the existing P. flaviventris, which, as we have seen, also occurs in the pleistocene; and by an unnamed species apparently more nearly allied to P. tectum and P. tentoria. Of Batagur, a genus confined to the Oriental region, there are four named species. Of these B. falconeri is regarded as the ancestor of the Indian B. thurgi; B. bakeri is equally closely allied to the existing B. kachuga; B. durandi shows a strong resemblance to B. dhongoka; while B. cautleyi presents affinity to B. affinis and B. pictus, respectively of the Malayan peninsula and Borneo. A generically undetermined nuchal scute from the Punjab may belong to a member of the genus Geoemyda. In the Trionychida, Emyda is represented by the existing Indian E. vittata, and also by three extinct species respectively named E. lineata, E. sivalensis, and E. palæindica. The three latter differ from existing species in the structure of the nuchal scute, and the last two are of comparatively large size. Trionyx was probably represented by several species, but none of the specimens yet obtained have afforded satisfactory characters for specific diagnosis. Lastly, the peculiar Indian genus Chitra is represented by the existing C. indica, which is the only known species.

In the Crocodilia, Crocodilus sivalensis is very closely allied to the existing Indian C. palustris; while C. palaindicus of Perim Island is another member of the same group distinguished by its convex facial profile. The gharials are represented by the existing Garialis gangeticus; and by an allied form known as G. hysudricus, which is distinguished by the form of the rostrum, and probably also by that of the cranium proper. G. curvirostris from the lower Siwaliks of Sind is a very distinct form characterised by the non-eversion of the anterior border of the orbit, and probably allied to certain upper cretaceous gharials of North America which have been generically separated under the name of Holops. G. leptodus is known only by the rostrum, which presents certain characters approximating it to Tomi

Quart. Journ. Geol. Soc. Vol. XLII. p. 540 (1886).

stoma. In G. pachyrhynchus of the lower Siwaliks of Sind we have a gigantic species provisionally referred to the type genus presenting certain marked peculiarities in the relations of the anterior teeth and their pits. Finally another gigantic species from the upper Siwaliks has been made the type of the genus Rhamphosuchus, with the specific affix of crassidens, and indicates a gharialoid presenting features in the relations of the upper and lower dentition now found only in the short-jawed alligatoroid members of the order.

The Lacertilia are only known by the gigantic Varanus sivalensis, which is estimated to have attained a length of eleven feet. The group appears to be an old one, as it is represented by the closely allied, if not generically identical, Palaovaranus of the Quercy phosphorites.

The only remains of Ophidia hitherto obtained from the Siwaliks are vertebræ of the genus Python from the Punjab and Sind. Some of these specimens have been provisionally identified with the existing Indian P. molurus, while others may indicate a distinct species. A python from the Quercy phosphorites, originally named Python cadurcensis, but regarded by some as generically distinct and therefore named Palæopython, carries back the origin of these huge serpents to a remote epoch. Eocene. The only specifically determined eocene reptile may be referred to the genus Platemys, under the name P. leithi. The specimen on which this determination rests is a carapace from the inter-trappeans of Bombay. The genus Platemys belongs to the Chelydida, and, although occurring in the Purbeck and lower eocene of England, is now confined to South America; the section or suborder (Pleurodira) of which that family is a member being characteristic at the present day of the southern hemisphere. From the nummulitics of the Punjab numerous fragmentary remains of crocodilians have been obtained, but are in too imperfect condition for determination. Remains of large chelonians have recently been obtained from below the coal-beds at Nila in the Salt-range, which are apparently transitional between the eocene and cretaceous. These specimens apparently belong to two genera; and for one which presents the peculiar feature of having the carapace covered with horny plates, while the plastron is pitted as in the Trionychida, the name of Hemichelys warthi is proposed; like Platemys, it belongs to the Pleurodira.

II. MESOZOIC AND PALEOZOIC.

Cretaceous. From the Arialur group (upper cretaceous) of the Trichinopoli series, and probably from the Lameta group (higher cretaceous), there have been obtained a few teeth of species of Megalosaurus, a genus whose range in England extends from the jurassic to the wealden, but is also found in the Maestricht beds of the Continent; the one tooth of the Indian form now forthcoming is in the Indian Museum. From the Lameta group there have also been obtained the remains of another genus of gigantic dinosaur, to which the name Titanosaurus has been assigned, which was represented by two species-T. indicus and T. blanfordi; the

1 Referred by Gray (Ann. Mag. Nat. Hist.' ser. 4, Vol. VIII., p. 339) to Hydraspis, which is now usually included in Platemys. The species was originally described by Carter as a Testudo.

former characterised by the centra of the caudal vertebræ being compressed, while in the latter they are sub-cylindrical. Numerous vertebræ, chiefly caudal, and a huge femur, nearly 4 feet in length, are preserved in the Indian Museum, and there is a cast of one of the former, belonging to T. indicus, in the British Museum. These forms, which have hitherto been regarded as allied to Ceteosaurus of the European wealden, but are referred to a separate family, appear much more closely related with Ornithopsis of the wealden, since a caudal vertebra from the Isle of Wight preserved in the British Museum, agrees very closely with the vertebræ of T. blanfordi, and agrees in relative size with Ornithopsis, of which the caudal vertebræ have been hitherto unknown. It may eventually prove that T. blanfordi is generically distinct from T. indicus. A few bones in the Indian Museum indicate a `smaller undetermined reptile from the Lametas. The Chelonia are known in the cretaceous by some broken plates, in the collection of the Indian Museum, obtained from the Lametas, from the infra-trappeans of Rajamahendri (Rajamundry), and from the upper cretaceous of Sind. The Crocodilia are represented by one amphicœlian species, apparently allied to Suchosaurus of the English wealden, of which some vertebræ have been obtained from the upper cretaceous of Sind, and are now in the Indian Museum. A large species of Ichthyosaurus named I. indicus is known solely by a few vertebræ obtained from the Utatur group (middle cretaceous) of the Trichinopoli series, and now in the Indian Museum; the range of the genus in Europe is from the lias to the chalk.

Gondwana system.-The exact identification of the different members of the great freshwater Gondwana system of Peninsular India with European horizons must be regarded as still unsettled, although it is probable that the fossiliferous beds range from the jura to the permian. From the Chari group of the jura of Kach there has been obtained a single vertebra, with an amphicœlian centrum, of which the affinities have not yet been determined, but which may be crocodilian 2; and from the Umia group of the same, a fragment of the mandible of a Plesiosaurus, described as P. indicus; the affinities of the latter form cannot be fully determined from the specimen.

The Maleri group, whose fauna agrees with that of the upper trias of Europe, has yielded the primitive crocodilian Parasuchus hislopi, which is apparently allied to Stagonolepis of the Elgin sandstone, and is the type of the suborder Parasuchia; Hyperodapedon huxleyi, differing from the triassic European H. gordoni by the arrangement of the palato-maxillary teeth, and apparently by the presence of lateral teeth on the mandible; and a dinosaur apparently allied to the European triassic genus Thecodontosaurus. The Tiki group in South Rewah, which is provisionally

1 In dealing with the vertebrates of certain Gondwana groups, I have shown that the evidence taken alone would indicate the following homotaxy of certain groups, vis. :—

(Jabalpur and Rajmahal.)

Low. Jura.:-Kota.

Up. Trias.:-Maleri.

Low. Trias.:-Panchet.

Up. Permian :-Bijori and Mangli (Upp. Damuda).

This would indicate that the Barakars (Low. Damuda) correspond either with the lower permian or the upper carboniferous, and the Talchirs either with the upper or lower carboni. ferous. See Appendix, and 'Records', Vol. XIX, pp. 133-34 (1886).

2 It has been suggested that this form may be Parasuchus; but the horizon renders this improbable.

с

correlated with the Maleri horizon has yielded the same rhynchocephalian and dinosaurian remains, but the crocodilian of the latter group is replaced by the European genus Belodon, so characteristic of the upper trias (keuper) of Germany. The oldest reptiles hitherto found in India occur near Ranigunj in lower Bengal, in the Panchet group of the Gondwanas, of which the vertebrate fauna has a lower triassic facies. The remains of a species of Dicynodon, belonging to the Ptychognathine group, are of comparatively common occurrence in the coarse Panchet sandstone, and have been described as D. orientalis. Other remains seem to indicate a second and larger species of the genus. The suborder of reptiles to which Dicynodon belongs is characteristic of the reputed trias of India, Russia, and Africa, and attained its fullest development in the latter country. The remains of the Indian forms all occur over a very small area in one thin seam of the Panchets. The dinosaur has been named Epicampodon & indicus, and is the sole representative of the genus; it is known merely by two minute, compressed, and trenchant teeth, with serrated edges like those of Megalosaurus, and implanted in distinct sockets. The above specimens are in the Indian Museum.

CLASS IV-AMPHIBIA.

I. TERTIARY AND POST-TERTIARY.

Pleistocene. Remains of a large Bufo, probably identical with the existing Indian B. melanostictus, have been obtained from the Karnul caves in Madras.

Eocene. In the eocene of Bombay there occur numerous remains of a small frog, belonging to the genus Oxyglossus, now living in China, Siam, and possibly India; the fossil species is extinct, and is known as O. pusillus: remains of a larger, but undetermined, frog are also indicated.

II.-MESOZOIC AND PALEOZOIC.

Gondwana system.-The Denwa group of the Satpura district, which is probably not far removed in time from the Maleri horizon, has yielded the right supratemporal bone of a large species of Mastodonsaurus closely allied to M. giganteus of the upper and middle trias (keuper and muschelkalk) of Europe. Remains of another large labyrinthodont, apparently allied to the upper triassic (keuper) Metopias and Capitosaurus, have been obtained from the Maleri and Tiki (South Rewah) groups; while the former group has also yielded fragments of a Pachygonia which may be specifically identical with the Panchet form.

From the Panchet group three genera of slender-jawed labyrinthodonts allied to those of the European lower trias are known. The first of these, Pachygonia, has only the one species P. incurvata, and is known by the greater part of the mandible, and a fragment of the cranium. The marking of the former is like that of Mastodonsaurus. The second genus, Gonioglyptus, has two species, the smaller known as G. longirostris and the larger as G. huxleyi; it is considered to be closely allied to Trematosaurus of the bunter-sandstone of Germany. The third genus is

1 The ordinal term Anomodontia is here used in a wide sense to embrace the Dicynodontia, Rhynchocephalia, Theriodontia, &c., which are ranked as suborders.

2 Originally described as Ankistrodon-a name preoccupied by a genus of Pisces (Ancistrodon).