1026 MR. D. M. S. WATSON ON 
"hese characters are, of course, of very different values, but, 
taken together, do show that the Bauridee differ enormously from 
the Cynognathide. 
It may be argued with an appearance of truth that these 
characters are all merely primitive ones, and that the Bauridee 
are derived from the same near ‘“ Cynodont” ancestor as the 
Cynognathids, but, although much specialised in some ways, are 
retarded in other features. This is in essence the view that 
Dr. Broom stated of Bauria in the form of a genealogical tree 
in December 1911, and which, I believe, he still holds. 
IT endeavour to show in the remainder of this paper that 
Dr. Broom’s view is now untenable. ‘The method I employ 
is to discuss, with the aid of a series of specimens, of which the 
relative ages ave known, the evolution of certain selected features 
of the skull of a Gynognathid, and then endeavour to show that 
the conditions in the Bauridee could not have been derived from 
those in any, except perhaps the earliest, of these stages. 
The series of forms | shall use are :-— 
Dimetrodon; Upper Carboniferous ¢ or Avtinskian. 
* Arctops willistoni, gen, et sp. n.; Hndothiodon zone? Middle 
Permian ¢ or earlier. 
Scymnognathus whaitsi? Broom; Hndothiodon zone. Middle 
Permian. 
Arctognathus curvimola (Owen); Cisticephalus zone. Upper 
Permian. 
Diademodon; Cynognathus zone. Middle Trias. 
The first of these is a North-American Pelycosaur, and the last 
a South-African Cynognathid. The other three are South-African 
Gorgonopsids. It is certain that they are not actual ancestors of 
one another, but they may be as nearly related as the animals 
included in the early phylogenies of the horse. 
Baur, Case, and v. Huene have published more or less complete 
accounts of the occiput of Dimetrodon, and I am indebted to the 
extreme kindness of Prof. 8. W. Williston for the gift of a magni- 
ficent specimen which has been of the utmost use to me. Dinetro- 
don has a large, round occipital condyle, in front of which the 
thick basioccipital forms the narrow floor of the brain-cavity, and 
articulates at the sides with the exoccipitals and opisthoties, 
which are usually distinct. The opisthotics form powerful par- 
occipital processes lying below the small post-temporal fossz. 
The small exoccipitals lie on the extreme posterior surface, 
projecting at the sides into processes which run outwards towards 
the post-temporal fosse far above the bottom edge of the par- 
occipital processes. The whole bone looks as if it had only 
recently become a part of the skull, and was still only very 
imperfectly connected with the rest of the occiput. The foramen 
jugulare lies high up at the back of the skull between the lower 
* Wounded ona skull in theG. G. Bain collection, B.M.N.H. R. 4099, from Howse’s 
Port. It is distinguished by the extreme width of the parietal region and the small 
size of the orbit. 
