OF THE MESOZOIC MAMMALIA. 
253 
the two to three row types is beautifully shown in Ghirox (p. 219). The molars of 
Stereognathus, which has provisionally been placed in this group, show an antero- 
posterior crescentic disposition, very similar to that observed in the transverse cres- 
cents of the primitive Selenodont Artiodactyle molars. The same is true of the 
Meniscotissus molars. The former is the only genus in which three rows of tubercles 
are found in the lower jaw. 
AVe can form no adequate conjecture as to the origin of the multitubercular 
molars. If the quadritubercular type, in other lines of descent, sprang from the 
single cone, there is no reason why the same should not have been the case here. 
There is some ground for this surmise, in the evolution of the multitubercular from 
the quadritubercular molars among the Plagiaulacidse, and the transition from the 
tri tubercular to quadritubercular premolars of the Bolodontidm. 
The reduction of the molars was evidently postero-anterior. The typical or 
stem dentition was probably i 3, cl, p 4, m 6. The hypothetical relationships of 
these families may be expressed in this diagram. The Plagiaulacidse and Bolodonti- 
dce seem to have diverged at an early period from one stem and the Tritylodontidae 
and Polymastodontidse from another. It is possible that the last two families were 
upon the same line. 
1 ^ 
P 
3 oi 
CO S. 
Neoplagiaulax 
^Hypothetical. 
