90 BULLP^TIN 16 9, UNITED STATES NATIONAL MUSEUM 



in another specimen (no. 9762) it runs backward into the maxilla for a 

 short distance and ends in a blind point. An alternative, and I 

 think slightly more probable, explanation is that this peculiar pocket 

 lodged a nasal diverticulum. The large paired palatal vacuities, 

 regularly long and elliptical in outline, extend from opposite the 

 posterior end of P- to the anterior end of M^*^ 



As shown vaguely in the best skull and definitely in no. 9710, the 

 maxillopalatine suture parallels the dental border and ends anteriorly 

 at the rim of the vacuity opposite the anterior end of M^ The 

 palatines thus form the whole of the quadrangular palatal bridge 

 principally between the first molars. In no. 9762 this is seen to be 

 pierced on each side by a long, large, horizontal canal opening anteri- 

 orly at the vacuity and posteriorly in the basicranial region, perhaps 

 in the choanae. The posterior palatal rim is slightly thickened. 



The choanae are completely separated by a thin, vertical, median 

 plate of bone, probably the vomer, as suggested by Broom. The 

 palatopterygoid crests are low and rounded and do not extend down- 

 ward to the level of the palate. There apparently was no hamular 

 process. 



The frontals are fairly large and cover most of the interorbital 

 region and form the superior border of the orbits. There are thin 

 lateral forward extensions of the parietals, which lap over the frontals, 

 as correctly shown by Broom, but they do not reach the nasals or 

 maxillae as in Taeniolabis. 



The anterior branches of the sagittal crest nearly follow the parieto- 

 frontal sutures, and the crest becomes single only near the junction 

 with occiput and is there low. The parietals are fused on the midline, 

 in distinction from the nasals and frontals. The presence of an inter- 

 parietal, as shov/n in broken lines by Broom, is purely hypothetical. 

 The parietosquamosal suture is not determinable. The large un- 

 broken piece of bone on the left side does not show it, and this suggests 

 that the posterior end of the suture was more lateral, and the parietal 

 here broader, than shown in Broom's restoration. 



The squamosal forms the posterior portion of the zygoma. Anterior 

 to the glenoid surface its lower face is slightly excavated, and this 

 may have been for the jugal, as suggested by Gidley and tentatively 

 shown on Broom's restoration, but this is by no means certain and 

 seems to me improbable. The glenoid surface, in any event, is wholly 

 on the squamosal and is oval and nearly plane. From it the squamosal 

 swings almost straight medially to the lambdoid crest. 



The basicranial region is very obscure, but a few details can be 

 made out. The anteroventral part of the occipital condyle is a very 

 thin flange underhanging a pocket in the posterior part of which is 



" They are considerably too short in Broom'srestoration, the anterior margin being placed too posteriorly. 



