84 MR. D. M. S. WATSON OX THE 



visible except the extreme articular edge and the two processes 

 which interlock with the squamosal. There is evidence that it 

 is no higher than the paroccipital process, but its upper edge and 

 front face are not seen. The bone extends only very slightly 

 laterally of the outer process, forming the extreme end of the 

 lower margin of the occiput. Above its end the border of the 

 squamosal runs upward and slightly outward, its front face being- 

 supported by the jugal. 



The plan of this very remarkable occiput is repeated in a less 

 exaggerated form in Scylacosaurus sclateri, so far as the still very 

 incomplete preparation of the British Museum skull allows of a 

 comparison. 



The brain-cavity of Scymnosaurus behind the epipterygoid is 

 now fairly well exposed. Its general characters will be best under- 

 stood from text-fig. 24. The foramen magnum is extremely 

 small, but the brain-cavity is somewhat larger than would be 

 anticipated ; although not high it is fairly broad, especially in 

 advance of the exit of the vagus. It is of normal Anomodont 

 type with an opening to the inner ear placed very low down and 

 with this opening confluent with the foramen for the Xth nerve. 

 The sunken edge which separates these openings is continued 

 upwards by a ridge on the wall of the brain-cavity, which 

 separates the narrow medullary from the wider cerebellar region. 

 The pituitary fossa, although not cleared of matrix, is undoubtedly 

 shallow. There is a powerful process below the notch for the 

 Vth nerve, and the supraoccipital, with possibly a strip of the 

 pro-otic, extends forward as side- walls as far as the epipterygoid, 

 passing median of that bone. 



The limb-bones doubtfully associated with the type-skull are 

 small and very slender, so that if they really belong to it the 

 proportions of the animal must have been like IJycenodon. If 

 they do belong the agility which they imply may be the ex- 

 planation of the unexpectedly large cerebellar cavity, which, 

 however, shows no trace of floccular fossae. 



It is interesting to compare Scymnosaurus with Scymncgnathus, 

 which is of about the same size, though later in time. The two 

 animals are carnivorous and have very similar dentition, especially 

 in the feeble molar series. 



The Therocephalian is the more advanced in the following 

 characters : — 



1. The reduction of all parts lying below the base of the 



brain, the basioccipital, basisphenoid, and especially 

 the quadrates. 



2. The reduction of the intertemporal region to a. narrow 



sagittal crest. 



3. The shortening of the snout. 



4. The lengthening of the temporal fossa?. 



Scymnosaurus rather recalls Scyrnnoyaathus in its square-cut 

 snout, Scylacosaurus is much more advanced in the rounding of 



