FAUNA OF THE KARNUL OAVES. 
31 
hind talon ; the cingulum is, however, absent in the first, although well-developed in 
the second Siwalik species, being in some instances 1 considerably larger than in the 
existing one. 
Canine . — An imperfect upper canine (No. F. 223) from bed Ce in the Cathedral 
may not improbably be referred to the present species. 
Origin of the species . — The occurrence of II. crocuta in the Karnul caves is of 
extreme importance in regard to the history of the species, since it connects its 
distributional area with that of the primitive crocutine Siwalik hyaenas, and leaves 
but little doubt that the species originated in India from H. colvini and thence spread 
westward into Europe at the close of the pliocene or commencement of the pleis- 
tocene epoch 2 (when it attained its largest dimensions), and subsequently reached 
Africa. The strongly-marked cingulum in the Karnul carnassial is especially note- 
worthy as indicating the close connection of this race with H. colvini rather than 
with H. felina. 
VlVERRA KARNULIENSIS, nobis. 3 
Definition . — The species may be defined as equal in size to V. zibetha, but 
distinguished by the much longer space occupied by the premolar series, in which 
respect it appears allied to the Siwalik V. baJceri. 
Mandible. — The type specimen of this species was obtained from bed L 
in tlie Charnel-House, 4 and is represented in pi. VII. figs. 6, 6a. The specimen 
consists of a fragment of the alveolar portion of the left ramus of the mandible 
showing the alveoli of all the premolars except pm. 1, and the complete mTI in a 
partially-worn condition. The latter tooth agrees very closely, both in size and 
structure, with the carnassial of Viverra zibetha ; almost the only observable difference 
being that the crown is relatively rather wider, more especially in its talon half. 
The premolar alveoli, as will be seen by the measurements, occupy, however, a 
much longer space, and the ramus when complete was evidently much deeper. In 
the following table the dimensions of the specimen are compared with those of the 
mandible of V. zibetha. 
V. karnuliensis. V. sibetha. 
Length of mTl 0'59 0-6 
Wid*h of do. at mandible 0'3 0*3 
Length of space occupied by alveoli of last three premolars . . . . 1’28 1-08 
Affinities . — Viverra megaspila agrees so closely with V. zibetha that what applies 
in one case applies in the other, and the lower carnassial of F. civetta is quite 
different from that of the fossil. None of the larger existing species of Herpestes 
agree with the latter, so that its distinctness from all living species may be taken for 
granted. Of the Siwalik species V. balceri 5 (of which the mandible is unfortunately 
unknown) agrees with the present form in relative size, as well as in the important 
character of pm. 3 being longer than in the existing species. 6 This probably indicates 
l Supra, vol. II. p. 296. fig. 14. 2 It occurs in the Norfolk Forest bed. 3 ‘Rec. Geol. Surv. Ind.’ vol. XIX. p. 120 (1886). 
4 See Foote * Rec. Geol. Surv. Ind.’ vol. XVII. p. 206. 6 Supra, vol. II. pi. XXXIII. fig. 1. 6 Ibid. p. 270. 
H 
