1028 MR. D. M. 8S. WATSON ON 
The basioccipital is a very thin bone, articulating as usual with 
the exoccipitals and opisthotics. 
The exoccipital is an exactly similar bone to that of Arctops. 
The paroccipital process is very shallow, so that the foramen 
jugulare lies much nearer the bottom of the skull, and actually 
looks as much downward as backward. The front face of the 
paroccipital process is slightly channelled, and the fenestra ovalis 
lies at the inner end of this groove above the level of the basi- 
occipital condyle. It is a comparatively small hole surrounded 
by a smooth and badly-preserved area of bone. 
The tubera are much lower and much further forward than in 
Arctops, the lower surface of the basisphenoid being only com- 
paratively slightly depressed. 
The basipterygoid process is very similar to that of Arctops, 
but seems to be pierced by a canalis Vidii. 
It is at once obvious that most of the differences between this 
type and Arctops are due to a reduction in depth of everything 
which lies below the bottom of the foramen magnum. 
The type and only existing skull of Arctognathus is unfortu- 
nately damaged so that nothing behind the fenestra ovalis is 
visible. In this form the basioccipital is even thinner than in 
Seymnognathus, and the fenestra ovalis lies above the bottom of 
the foramen magnum. The tubera basisphenoidalia are repre- 
sented only by the thickened margins of the slightly concave 
triangular lower surface. The basipterygoid process is similar to 
that in Seymnognathus, but is relatively further forward. 
In Diademodon the basioccipital is so thinned that the centre 
part of the thin lip-like condyle of Seymnognathus is pinched out 
altogether, leaving the two condyles, which are no doubt mainly 
exoccipital. Although no definite suture has ever been seen, the 
texture of the bone makes it certain that the exoccipital in 
Diademodon is a small triangular bone forming the lower margin 
on the extreme back of the skull. The fenestra jugularis les 
entirely on the lower surface and does not face in the least 
backward. The paroccipital process is more massive than in 
Scymnognathus owing to the fact that the post-temporal fosse lie 
above the level of the foramen magnum. ‘The front face is 
channelled, and the fenestra ovalis les at the inner end of the 
groove considerably above the lower margin of the foramen 
magnum. 
The tubera basisphenoidalia are very reduced, being merely 
the thick edges of the nearly flat ventral surtace of the triangular 
basisphenoid. 
The horizontal basipterygoid process is similar to that of 
Scymnognathus, but much further forward. 
Arctognathus is in all visible features of this region exactly 
intermediate between Scymnognathus and Diademodon. 
Comparison of these descriptions with text-figs. 3 and 4 will 
show that the changes take place quite regularly with time, and 
that all of them depend eventually on the reduction of the 
whole regions which lie below the base of the brain, a reduction 
