SIWALIK AND NARBADA PROBOSCIDIA. 
37—218 
length. The medial line of this rostrum is deeply hollowed out posteriorly, the hollow 
gradually diminishing anteriorly ; its depth posteriorly, where the specimen is broken 
off, is upwards of 7 inches, the lateral walls are thin, and nearly vertical externally. 
There are no traces of tusks in this specimen. There must have been an interval of 
at least a foot between the proximal extremity of this symphysis and the fragment 
of the ramus of the mandible where broken off in front of the penultimate 
molar. 
Tushed symphysis. — We have now shown that Mastodon pandionis had a pro- 
duced symphysial rostrum, which in some individuals was tuskkss ; this leads us to 
the consideration of another mandibular symphysis (Plate XXXVI, fig. 2), which 
from its form evidently belonged to a Mastodon^ and I think most probably to M. 
pandionis. This specimen was, like the other, obtained by Mr. Theobald in the Punjab. 
It consists of the distal extremity of the symphysis, the fragment being about one 
foot in length. It is furnished with a pair of large and laterally compressed tusks, 
of which the middle portions alone remain. These tusks are strongly arcuated, their 
superior border being concave and their inferior convex : they consist entirely 
of ivory, and present no trace of any band of enamel, as seems to be uni. 
versally the case in the genus. Their transverse diameter is 2 inches, and the 
vertical 3'1 inches. The rostrum itself is concave superiorly and convex inferiorly 
in the antero-posterior direction. The anterior portion of the superior surface 
presents a lozenge-shaped hollow, externally to which this surface slopes away to 
join the lateral surfaces. In this specimen there is no trace of the deep longitudinal 
groove which occurs in the tuskless symphysis represented in fig. 1 of the same 
plate. Such a groove could not, however, exist between the large tusks of the 
second symphysis, and if, as I think is most probably the case, the latter belonged 
to the same species as the former, it would seem that the presence of the large tusks, 
in the one and their absence in the other are the cause of the differences in form of 
the two specimens. The symphysis of the mandible is known in Mastodon latidens, 
M. perimensis, and M. sicalejisis, as we shall see when we come to the description of 
those species, and the tusked symphysis cannot, therefore, belong to either of them, 
all of which have much shorter symphyses. The complete mandible of 31. falconer i 
is unknown, but the specimen already referred to (Plate XXXIII, fig. 1) proves 
that this species had a short and probably edentulous symphysis. If, therefore, 
the tusked symphysis under consideration does not belong to M. pandionis, it must 
belong to a new species of Siwalik Mastodon, which is somewhat improbable. On 
the supposition that the symphysis in question belongs to M. pandionis, it must 
have belonged in all probability to a male individual, while the tuskless specimen 
represented in the other figure of the same plate must have probably belonged to a 
female. 
Last upper true molar. — A last left upper true molar of a trilophodont Mastodon 
in the Indian Museum seems to belong to the present speeies. The specimen was 
collected by Mr, Theobald in the Punjab ; but is not in a good condition for fi guring 
