1878. ] 529 [Cope. 
aa, The shaft interrupted by a prominent diagonal ridge. 
No, 4 -Epicondyles and ridges Enormous. <2... 6s. woe ee ne cress es Ps le 
AAA, No supracondylar fossa ; condyles as in 4 4, 
INO. Oe MODnh MOLE SION Grsa. accretive -ciivecos tats so sae asa. De O- 
IN O26). HORMEMOLe TOD Ustentets av-iceiereiteertealeters cistaciatels/erre sie sia ecole Dak 
The above humeri represent three, and perhaps four genera, which have 
been probably already named from crania or vertebre in the preceding 
pages. No. 1 has been already identified as belonging to the Clepsydrops 
natalis. Nos. 2 and 3 are generally similar to the type referred by Meyer 
to the Hurosaurus of Fischer, which had been previously described as 
Mammalian by Kutorga ; but the epicondyles are more largely developed. 
Humerus of form No. 4 is very remarkable, resembling in some degree 
that of a mole, being exceedingly robust, and having the muscular inser- 
tions enormously developed. It doubtless belonged to a fossorial animal, 
possessing great power in the anterior limbs. If we search for vertebra 
presenting features corresponding to such a mode of life, we sieze at once 
on those of the genus Hmpedocles. Here the elevated roof-like character 
of the zygapophyses and the connecting platform suggest protection against 
superincumbent weight, while additional strength is obtained by the 
hyposphen articulation below them. The short wide neural spine is high- 
ly appropriate also to subterranean habits. It is also probable that the 
animals possessing the humeri, from No. 2 to No. 6 inclusive, were all 
more or less fossorial. Humeri Nos. 5 and 6 have the characters of Nos. 
2 and 3, but the supracondylar bridge is wanting, and the internal epicon- 
dyle not quite so much expanded. 
The division Pelycosauria is established primarily on the genera Clepsy- 
drops and Dimetrodon, but their cranial structure renders it highly proba- 
ble that Hetocynodon, Pariotichus and Bolosaurus belong to it. It is also 
probable that the genera Hmpedocles, Hmbolophorus and others determined 
from vertebrae belong to it, as the latter are frequently accompanied by 
pelvic bones of the type of that of Dimetrodon. All the genera known 
from teeth ard crania, are of carnivorous habit, excepting Bolosaurus and 
Diadectes ; they may be referred to a single family on this account, which 
I call the Clepsydropide. Bolosaurus will form the type of another fam- 
ily characterized by the transverse position of the crowns of the teeth, un- 
der the name of Bolosauride. Prof. Owen has named a group of Triassic 
and Permian reptiles the Zheriodonta, characterized by the mammal-like 
differentiation of the incisor and canine teeth. The animals thus referred 
by Prof. Owen probably enter my suborder of Pelycosauria, although the 
structure of their pelvis remains to be ascertained. Ifso, they correspond 
with my Clepsydropide, since Prof. Owen does not include herbivorous 
forms in his division. As itis plain that the herbivorous and carnivorous 
types belong to the same order, and probably suborder, it becomes neces- 
sary to subordinate the term Zheriodonta to that of Pelycosauria. To 
another division of reptiles from the South African Trias typified by the 
PROC. AMER. PHILOS. soc. XvII. 101. 3N. PRINTED MAy 18, 1878. 
