e modifications of the quadrate-articular
joint, and the development of the coronoid process.

Thrinaxodon

The evidence for the position and extent of the external adductors of
the lower jaw in _Thrinaxodon_ was secured in part from dissections of
_Didelphis marsupialis_, the Virginia opossum. Moreover, comparison of
the two genera reveals striking similarities in the shape and spatial
relationships of the external adductors. These are compared below in
some detail.

The sagittal crest in _Thrinaxodon_ is present but low. It arises
immediately in front of the pineal foramen from the confluence of
bilateral ridges that extend posteriorly and medially from the base of
the postorbital bars. The crest diverges around the foramen, reunites
immediately behind it, and continues posteriorly to its junction with
the supraoccipital crest (Estes, 1961).

In _Didelphis_ the sagittal crest is high and dorsally convex in
lateral aspect, arising posterior to and medial to the orbits, reaching
its greatest height near the midpoint, and sloping down to its
termination at the supraoccipital crest. Two low ridges extend
posteriorly from the postorbital process to the anterior end of the
sagittal crest and correspond to ridges in similar position in
_Thrinaxodon_.

The supraoccipital crest flares upward to a considerable extent in
_Thrinaxodon_ and slopes posteriorly from the skull-roof proper. The
crest extends on either side downward to its confluence with the
zygomatic bar. The area of the crest that is associated with the
temporal musculature is similarly shaped in _Didelphis_.

The zygomatic bar in each genus is stout, laterally compressed, and
dorsally convex on both upper and lower margins. At the back of the
orbit of _Thrinaxodon_, the postorbital process of the jugal extends
posterodorsally. At this position in _Didelphis_, there is but a minor
upward curvature of the margin of the bar.

In _Thrinaxodon_ the dorsal and ventral postorbital processes, arising
from the postorbital and jugal bones respectively, nearly meet but
remain separate. The orbit is not completely walled off from the
adductor chamber. The corresponding processes in _Didelphis_ are
rudimentary so that the confluence of the orbit and the adductor
chamber is complete.

The adductor chamber dorsally occupies slightly less than half of the
total length of the skull of _Thrinaxodon_; in _Didelphis_ the dorsal
length of the chamber is approximately half