16 INDIAN TERTIARY AND POST-TERTIARY VERTEBRATA. 
the two specimens are from the same side of tlie jaw, they are excellently adapted 
for exhibiting the characteristic differences. 
As the result of the foregoing comparisons, the present writer (although fully 
aware of the difficulty of distinguishing the horses solely by means of their 
dentition) is very strongly of opinion that the teeth under consideration indicate a 
third Siwalik Ilippotherium, intermediate in size between the other two species. It 
was shown in the second volume^ that a lower jaw from the Punjab provisionally 
referred to U. theohaldi differed in some respects from the type specimen, and it is 
possible that it may belong to the same form as the upper molars from Perim. 
The latter are distinguished from the upper teeth of H. gracile in much the 
same respects as from those of H. antilopinum. It ai^pears, moreover, tliat in all the 
Indian species the plications of the dentine and enamel are decidedly more complex 
than in the European species ; this character being especially noticeable in well- 
worn teeth.'^ 
The writer has been unable to identify the Perim teeth with any of the 
American species. They are markedly distinct from Jl. calainariimi^ in which the 
hinder pillar of the upper cheek-teeth is remarkably large ; and they differ from 
H. speciosum^ by their superior size, by the more cylindrical anterior pillar, and the 
greater com25lexity of the plications of the enamel. 
Specific distinctness . — If further discoveries confirm the conclusion that the Perim 
teeth probably belong to a new species, tlie name H. feddeni may be ap]3ropriately 
applied. 
Affinities of the Suoalik species . — The extreme complexity of structure of the 
molars of the Siwalik hippotheres, coupled Avith the absence of pni. i in at least one 
species, points to the conclusion that none of these species were on the direct 
ancestral line of Equus. 
SuB-OuDER : PROBOSCIDIA. 
Nomenclature of the milk and premolar series . — In the first volume of this work,® 
owing to the general absence in the Proboscidia of the first deciduous cheek-tooth of 
the typical eutherian series, the three teeth of this series which are normally 
developed in the elephantine family of that sub-order were, folloAving Dr. Falconer, 
respectively termed antepenultimate, penultimate, and last, or more generally, first, 
second, and third. Although this nomenclature is convenient for the Proboscidia 
when considered by themselves, it is apt to lead to confusion Avhen treating of the 
other sub-orders of the Ungulata,® and the rest of the Mammalia ; and it, therefore, 
seems best to adopt the nomenclature of the typical eutherian series. In this 
1 Page 20, pi. XII., fig. 4. 
2 Compare the specimens figured in Gaudry’s “ Ann. Foss, et Geol. de I’Attique,” pi. XXXIV., fig. 7. 
3 Cope, “Rep. U. S. Geog. Surv. W. of 100th Meridian,’’ vol. IV., pt. II., pi. LXXA^., fig. 1. This species is 
distinguished from the European and both the other Indian species hj’' the want of the anterior projection of pm. 2. 
4 Ibid, fig 3. 5 Pp. 198-201. 
6 The discrepancy was not so marked when the Proboscidia were regarded as a group of equivalent value with the 
Ungulata. 
