70 Bihliographical Notice. 



The remains next described are such that no attempt is made to 

 characterize the genus to which they belong ; and the species are 

 indicated by numbers. Number one is about half as large again as 

 Testudo eUphantina, and is known from its epiplastron, which is 

 about intermediate in character between that of Colossoclielys atlas 

 and the epiplastron of Manotu-ia emijs. A second species, also indi- 

 cated by an epiplastron, is about one fourth larger than Testudo 

 elephantina. To this type the marginal jDlate described as CautJei/a 

 annxdiger may possibly be referred. Tiie third species is another 

 Punjab form, only known from the epiplastron. The fourth species, 

 which is rather smaller than Testudo eleplmntina, is known from 

 the epiplastron and one or two other bones, including the niTchal 

 scute. Unlike the species one, two, and three, this type is related 

 to the existing land-tortoises of India and Burma. 



The author then treats of seven species which are referred to the 

 genus Clemmys. They are arranged as having — first, no keel on 

 the carapace of the adult ; secondly, as marked by three continuous 

 keels ; and third, with three interrupted keels. Under tlio first head 

 are placed Clernimjs sivalensis (Theobald), C. liydaspica, C. Theo- 

 baldi, and C. punjahiensis. Clemmys hydaspica is closely allied to 

 C, sivalensis, from which it appears to differ chiefly in the form of 

 the first vertebral shield. C. TlieohaJdi shows the first vertebral 

 shield, remarkable for the prolongation of the anterior angles of its 

 pentagonal form. The author remarks that there are indications 

 that costal keels, which have become practicall}^ obliterated, existed 

 at an early age. C. punjahiensis is remarkable for the subquadrate 

 form of the vertebral plates. It most closely resembles the North- 

 American C. ventricosa ; but the Indian types all have the carapace 

 more elevated than any of the North- American species. The fifth 

 species is not named, but stated to be allied to OJemmys trijuga ; it 

 has three indistinct obtuse keels, and is distinguished by the form of 

 the vertebral plates. Clemmys palmndiea has three interrupted 

 keels, and shows a general resemblance to C. HataiUoni, which is 

 found in Lower Bengal. In the living form the vertebral plates 

 are wider than long; but in the fossil their relative length is 

 greater. 



The genus PangsJmra is represented both in the Narbada and 

 Siwalik beds by the P. jiaviventris (Giinther). A second species, 

 numbered 2, from the Siwaliks, is regarded as being related to P. 

 tentoria and P. tectum. Five species of the genus Batagur are 

 described. One, Batagur Falconeri, is allied to the B. thurgi of 

 Gray ; B. Balceri is allied to B, Icaclmga (Gray) ; B. Durandi is 

 allied to B. dhom/olrt (Gray) ; and in all these cases the fossil is 

 regarded as the ancestor of the existing form. The fourth species 

 is imperfectly known, and the fifth, B. Cautleyi, is distinguished by 

 the shortness of the fourth vertebral plate, and it differs from 

 B. Burundi, with which the author compares it. It is most nearly 

 related to the living B. affinis and B. pictus. A fragment showing 

 a triangular nuchal scute is referred to the genus Geoemyda. 



The Trionychidfe are represented by three genera — Enn/da, 



