CLASSIFICATION OF THE THERIODONTIA. 73 



Varanosaurus and Dimetrodon in dentition, as Case has already 

 shown. 



In the palate Dimetrodon diners from Varanosaurus in that the 

 transverse flanges of the pterygoids have moved much forward 

 from the basipterygoid region. The condition in Deiopeus is not 

 known. Thus many of the structural changes which separate the 

 advanced from the primitive Pelycosaurs are in the same direction 

 as the general advances of the Anomodonts from first to last. 



The earliest known Gorgonopsid is Galesuchus gracilis Haugh- 

 ton, from the Tapinocephaluszone. 



This skull I have not seen, but judge that it is so much 

 weathered that the original shape of the squamosals cannot be 

 seen. Only the dorsal aspect has been figured. The skull is very 

 remarkable, differing much from all later Gorgonopsids and even 

 more from Pelycosaurs. Its only primitive features seem to be 

 the extremely sloping occiput (the apparent slope being possibly 

 exaggerated), the high position of the foramen magnum and 

 foramen jugulare, the large deep paroccipital processes, the 

 great size of the frontal, and the lateral direction of the orbits. 

 It is, however, so incompletely known that a full discussion of 

 its affinities is impossible. 



The most primitive known Gorgonopsid is Arctops. This re- 

 tains as primitive features, which it shares with the Pelycosaurs: — 

 the square section of the snout with a, depression on the preorbital 

 surface overhung by the prefrontal, the lateral direction of the 

 orbits, the extreme parietal width, and the shortness of the 

 parietals. The resemblances in the basicranial and otic regions 

 have been discussed (P. Z. S. 1914, p. 1027). 



The palate retains a primitive structure in the non-fusion of 

 the prevomers. 



The skull shows advances over Varanosaurus in the following 

 ways : — Owing to the need of increasing the size of the temporal 

 muscles the squamosal is bayed outward, a modification which 

 makes the temporal fossa visible from above ; this change not 

 only allows of greater thickening of the temporal muscles during 

 their contraction, a function which is believed by Gregory and 

 Adams to be the factor which determines the origin of fenes- 

 tration, but also enables the outer part of the temporal muscle to 

 acquire a new origin on the upper edge and inner surface of the 

 zygomatic arch, thereby establishing an independent masseter 

 muscle. At the same time this widespread zygomatic arch passes 

 much laterally to the quadrate and leaves that bone, with 

 the quadrato-jugal attached to its outer margin, lying entirely 

 within the back of the enlarged temporal fossa ; these bones, 

 having thus lost the support that they originally received from 

 the junction of the quadrato-jugal with the jugal and squamosal 

 by sutures, can only be adequately supported by a more powerful 

 abutment on the paroccipital process and by an extension of the 

 squamosal down their posterior surface. 



At the same time the lateral extension of the squamosal renders 



