SIWALIK AND NARBADA CHELONIA. 
19—173 
characteristic of C. sivalensis. The difference in the form of the shells is indicated 
by their respective dimensions ; the width of the present specimen being 5T inches, 
and the height of the carapace 2’6 inches. The two forms agree in the absence of 
any trace of either vertebral or costal keels,, and of areolae. In profile (fig. 4a) the 
contour of the carapace of the present form is very similar at its two extremities. 
It cannot, unfortunately, be determined whether the present form possessed a 
nuchal plate : the anterior marginals are relatively wide antero-posteriorly. The 1st 
vertebral plate closely resembles that of C. sivalensis, with the exception that its 
posterior border is entire. The 2nd and 3rd vertebrals also have a strong general 
resemblance ; but their anterior borders in place of being sinuated with a median 
projection, are convex and continuous with the anterior lateral borders, so that these 
plates become more ’’ balloon-shaped,’^ and are practically three-sided : their posterior 
borders are entire. The 5th vertebral is extremely wide ; its inferior border 
equaling the united width of the pygal and last marginal plates. The gular and 
postgular plates are apparently very similar to those of C. sivalensis. 
Distinctness and affinities . — The foregoing comparisons indicate pretty clearly the 
specific distinctness of the present form from G. sivalensis^ ; but the resemblance in 
the shape of their vertebral plates probably indicates their affinity. C. crassicollis is 
distinguished from the present form by its more depressed carapace, less ‘ balloon- 
shaped ’ middle vertebrals, much narrower 5th vertebral, shorter gular plates, and 
the straighter lateral edges of the carapace. Whether the two agreed in possessing a 
nuchal plate, and in having serrated posterior marginals cannot be determined from 
the present specimen. As no other species with which the writer is acquainted^ 
presents any close resemblance to it, the fossil may be regarded as specifically 
distinct, and it is proposed to name it Clemnvjs hydaspica, from the classical name of 
the river Jhelam. 
The writer does not think it probable that C. hydaspica is the direct ancestor of 
G. crassicollis, but in connection with the preceding and succeeding species it indicates 
the great richness in forms which the group of emydines to which that species 
belonged attained in the pliocene of India. 
Species 3. Clemmys theobaldi, n. sp. nohis. 
Shell. — This species is founded on the shell represented of one-half the natural 
size in plate XX. figs. 2. 2a. 2b.; which was obtained by Mr. Theobald from the 
Siwaliks of Jhand in the Punjab. It is perfect as far back as a short distance 
behind the 3rd vertebral plate, where it is broken off by a fracture extending 
through the carapace and plastron. The greater part of the plastron is preserved, 
1 To Anglo-Indian readers an inverted ‘ gharra ’ would best convey the form of these plates. 
2 In certain tortoises, e.q. Testudo elongata, consideiahle differences in the form of the carapace of the two sexes is 
observable, but nothing comparable to that existing between U. sivalensis and the present form, C. cvassioUis, moreover, does 
not present such differences ; and there is also the probability that the small specimens noticed above belong to females of 
0. sivalensis. 
3 With rcSgard to CUnimgs nuvhalis, see the next species. 
