CLASSIFICATION OF THE THERIODONTIA. 69 



the lower surface of tlie basipterygoid process, and there is no 

 quftdrate ramus, the bone terminating at its attachment. The 

 transverse ramus does not arise from the basipterj'goid region, 

 but very much further forward, the posterior part of tlie bone 

 being a slender strip, which is united with its felloAV and the 

 parasphenoid to form a slendei' )jar. 



The transverse ramus arises very suddenly from this bar, the 

 bone being drawn downward into deep and very powei-ful flanges. 



There is no anterior lamus, the bone ending in a transverse 

 suture with the palatine at the region where ic joins the jugal. 



The transverse bone is very small and is not included in the 

 flange at the foot of which it lies. 



The palatine is a very large bone, forming a great deal of the 

 posterior part of the palate, then turning downward to form the 

 lateral wall of the posterior part of the nasopharyngeal duct, 

 and, finally, developing a secondaiy plate, which forms a floor to 

 that passage. 



The two pala,tines are separated by a median vomer, which 

 forms the roof of the nasopharyngeal passage, apparently ter- 

 minating behind in a pointed slip separating the anterior ends of 

 the pterygoids, but really passing backward to the basisphenoid. 



The prevomers seem to have vanished entirely. 



The maxilla? send inward secondary plates, wdiich continue those 

 of the palatine forward. The premaxillas have pala,tal processes, 

 which pass dorsally to the secondary plates of the maxillae and 

 touch the anterior end of an ossiScation in the nasal septum, 

 viz., a mesethmoid. 



In VaranosauriijS the epipteiygoid is a slender rod rising 

 vertically fronr the dorsal surface of the pterygoid just in advance 

 of the basipterygoid articulation. 



In Diademodon the epipterygoid is a large flat sheet of bone 

 forming a side-wall to the br.iin-case and articulating by a very 

 long suture with the parietal and fi'ontal above, having a sutuie 

 with the pro-otic behind, articulating with the basipterygoid 

 process below, and ending in a suture with the pterygoid. 

 Behind the basipterygoid and belo\A- the j^oint of exit of the 

 maxillary, mandibular, and motor portions of the fifth nerve, the 

 epipterygoid is continued backwaid by a process occupying the 

 position of a quadrate ramus of the pterygoid. The ramus in 

 certain species articulates with the front face of the quadrate. 



The preceding pages record the more imjDortant differences 

 between the most primitive and the most advanced known 

 Anomodont ; they bring out the direction of the evolutionary 

 advances and show how enormous is the structural gap between 

 them, a gap represented in time by the relativel}' small interval 

 between the bottom of the Permian and the middle of the 

 Triassic system. 



It remains to show how completely this morphological gap can 

 be bridged by the material available. Although, as Case and 

 Williston have repeatedly and emphatically pointed out, the Pely- 

 cosauvs are a self-contained group dying out with Dimetrodon and 



