th medially, thereby
producing a small depression at the junction of each base with the
dentary bone.

The lateral compression of the teeth is pronounced but asymmetrical, in
that the lateral surface of each blade is more convex than the medial
surface.

[Illustration: FIGURES 4-6. _Thrausmosaurus serratidens_, lower Permian,
6 miles north of Fort Sill, Comanche County, Oklahoma. All x 3.

FIG. 4. KU 11120 (type specimen), lateral view of left dentary.
FIG. 5. KU 11121, lateral view of ?left maxilla.
FIG. 6. KU 11122, lateral view of left dentary.]

The recurvature of the anterior cutting edges is much more severe than
that of the posterior edges, but the recurvature of both is limited to
the distal half of each tooth.

The serrations of the cutting edges are not visible to the naked eye and
are limited on the anterior edges of the teeth to those portions of the
blades that are recurved. The posterior serrations extend nearly to the
junction of the blade of each tooth with its base. The serrations tend
to be more nearly crenulate than cuspidate.

A portion of the lateral wall of the dentary surrounding the Meckelian
canal is present. The external surface of the wall is gently convex and
smooth, without sculpturing. The internal surfaces of the canal are
unmarked either by muscle scars or foramina.

The fragment is a piece from the posterior portion of the dentary, since
the decrease in height from the first tooth to the fourth is pronounced.

KU 11122, a fragment of the left dentary bearing two teeth, is 7.5 mm.
long. The anterior tooth is 3.0 mm. long; the posterior tooth is 3.5 mm.
long. The shape of the teeth and their implantation conform to the
description of the type specimen. The lateral surface of the fragment is
smooth and gently convex. What little is present of the surface
bordering the Meckelian canal is unmarked.

The ?maxillary fragment bears two teeth which are 3.0 mm. long, and
which conform in their characters to the type. The lateral, medial and
ventral surfaces of the fragment have been sheared off, so that an exact
identification of the bone is impossible. Presumably the fragment is too
deep dorsoventrally to be a piece of the dentary, and no sign of the
Meckelian canal is present.

_Discussion._--The implantation, lateral compression, recurvature and
cutting edges of the teeth borne by these fragments make clear their
sphenacodontid nature. The characters of the fragments are too few to