spicuous modification in _Captorhinus_ is the absence of the
tabular. This element in _Protorothyris_ was limited to the occiput and
rested without sutural attachment upon the squamosal (Watson,
1954:338); later loss of the tabular could have had no effect upon the
origins of muscles from inside the skull roof. Changes in pattern that
may have modified the origin of the adductors in _Captorhinus_ were
correlated with the increase in length of the parietals and the
reduction of the supratemporals. Other changes that were related to the
departure from the primitive romeriid condition of the adductors
included the development of a coronoid process, the flattening of the
quadrate-articular joint, and the development of the peculiar dentition
of _Captorhinus_.

The adductor chambers of _Captorhinus_ are large. They are covered
dorsally and laterally by the parietal, squamosal, postfrontal,
postorbital, quadratojugal and jugal bones. The chamber extends
medially to the braincase, but is not limited anteriorly by a bony
wall. The occiput provides the posterior limit. The greater part of the
adductor chambers lies mediad of the mandibles and thus of the
Meckelian fossae; consequently the muscles that arise from the dermal
roof pass downward and outward to their insertion on the mandibular
rami.

_Mandible_

The mandibular rami of _Captorhinus_ are strongly constructed. Each
ramus is slightly convex in lateral outline. Approximately the anterior
half of each ramus lies beneath the tooth-row. This half is roughly
wedge-shaped in its lateral aspect, reaching its greatest height
beneath the short posterior teeth.

The posterior half of each ramus is not directly involved in supporting
the teeth, but is associated with the adductor musculature and the
articulation of the ramus with the quadrate. The ventral margin of this
part of the ramus curves dorsally in a gentle arc that terminates
posteriorly at the base of the retroarticular process. The dorsal
margin in contrast sweeps sharply upward behind the teeth and continues
posteriorly in a long, low, truncated coronoid process.

A prominent coronoid process is not found among the more primitive
members of the suborder, such as _Limnoscelis_, although the mandible
commonly curves upward behind the tooth-row in that genus. This area in
_Limnoscelis_ is overlapped by the cheek when the jaw is fully adducted
(Romer, 1956:494, Fig. 213), thereby foreshadowing the more extreme
con