1889.] The Proboscidia. 197 
of the species referred above to Mastodon, mandibular tusks 
are present in the young, and occasionally one is retained to 
dividuals are exceptional among their species. In some 
other species while the males possess them, they are wanting 
to the females. The specific character is in this case derived 
from the male. 
The molar dentition in this family possesses a number of 
peculiarities which have been worked out mainly by Falconer, 
Owen, and Lydekker. There are probably deciduous molars 
in all the species, and they are generally three in number. 
The posterior of these has the same number of cross-crests 
as the posterior premolar, which immediately succeeds it. 
The number of crests diminishes to the first of the series. 
There are two or three premolars in most forms of the family, 
but in the genus Elephas they have disappeared. In all the 
species they are shed early in life in order to make way "for the 
true molars. As the latter teeth are very large, and the fore 
and aft extent of the jaws is small, there is only space for one 
or two of them at a time. In most of the species the last 
molar so much exceeds the others in size, that it occupies the 
entire jaw, and the other molars are shed in order to accom- 
modate it. In the genera Tetrabelodon, Dibelodon, and 
Mastodon, the last premolar, and the first and second true mo- 
lars are isomerous, /. e. have the same number of cross-crests. 
In Emmenodon and Elephas they are heteromerous ; that is, 
the number of cross-crests successively increases from front 
to rear. Thus in the three genera named the ridge formula 
is ; P. M. 2—2—3 ; M. 3—3—4, and P. M. }—> 4 I M. 4—4— 
5 or 4 — 5 — 6. In Emmenodon the ridge formula is, P. M. 
•'—?—?— 5 ; M. 6-'j-<>~7-Z ; and P. M. }-6-7 ; M. 7—^- 
9— 10-12. In Elephas the formula extends from M .6—6- 
7 — 8-9, to M. 9-15 — 14-16 — 18-27. Each genus then has 
a certain range of variation in the number of molar crests, 
extending from a smaller to a larger number. This succes- 
sive increase in complexity has been regarded by Falconer as 
the index to the successive evolution of the species, and 
rightly so. As already remarked, however, other measures 
of the same succession cannot be overlooked, especially as 
