70 MR. D. M. S. WATSON ON THE 



Edaphosaurus , they show evolutionary changes which are in the 

 main in the same direction as those which lead from Varanosaurus 

 to Diademodon, coupled, of course, with many individual speciali- 

 zations. Although it is not essential for the purpose of this 

 paper, it is, I think, useful to point out certain of the more 

 striking of these advances, using Varanosaurus, Deiopeus, and 

 Dimetrodon as the series of forms. These animals lived side by 

 side, and this series is only a morphological one. 



In general shape Dimetrodon shows an advance on Varano- 

 saurus in that the snout is much deepened, is square in section 

 only immediately in front of the orbit, and is, in general, wedge- 

 shaped with a rounded dorsal edge. The parietal region is 

 narrower, and leaves the temporal fossae visible from above. The 

 occiput is more vertical. In the brain-case Dimetrodon shows an 

 advance over Deiope%is in that the basiocoipital is thinner, the 

 basisphenoidal tubera smaller, and the paroccipital process larger. 

 The anterior margin of the pro-otic lies further in advance of the 

 internal auditory meatus. The brain-cavity is considerably 

 deeper and wider posteriorly. The fenestra, ovalis of Dimetrodon 

 is smaller than that of Deiopeus and the stapes lighter. The roof 

 of the skull of Dimetrodon differs from Varanosaurus in the 

 following ways : — the parietals are less wide and the pineal 

 foramen further forward. The postorbital is visible from above. 

 The pair of frontals have acquired a cruciform shape owing to a 

 widening of the interorbital surface. Deiopeus provides an exact 

 intermediate, the increased width of the interorbital surface 

 having arisen by an increase in size of the pre- and post-frontals, 

 so as to leave a gap which is filled up by a special process of the 

 frontal . 



In side-view the skull of Dimetrodon shows a shortened and 

 deepened maxilla touching the nasal in a short suture. The 

 dentition is sharply divided into incisors and cheek-teeth by a 

 diastema. The third maxillary tooth is much larger than the 

 first two, and the lower border of its socket lies well below the 

 dentigerous border of the premaxilla. 



The lachrymal does not reach the septomaxilla. The orbit is 

 placed high up in the skull. 



It seems probable that the deepening of the maxilla and the 

 " step " depend on the necessity of finding room for the roots of, 

 and the development of, replacing teeth for the greatly lengthened 

 maxillary teeth. 



The condition of the canine may depend on the following con- 

 siderations : — A large canine in the upper jaw presupposes a 

 similar tooth in the lower jaw ; such teeth, which are designed 

 for killing animals, are most useful in the front of the mouth. The 

 lower jaw, as a whole, bites inside the upper jaw. The first tooth 

 of the lower jaw cannot be much enlarged, because of the difficulty 

 of making a pit for its reception near the middle line in the 

 palatal process of the premaxilla ; hence the lower canine cannot 

 be quite at the end of the jaw. A large lower canine almost 

 involves a diastema in the upper jaw for it to bite into, because 



