782 ; MR. D. M. S. WATSON ON 
vertically in the skull, and les at the lower and inner end of the 
groove which Broom and ‘I, following the original identification 
of W. K. Gregory, regard as an external auditory meatus. If 
this position was really that held by the tympanic membrane in 
Diademodon, then in that type the tympano-tubal cavity had 
already grown up round the bones of the back of the lower jaw. 
{f this position of the membrane be correct, by tracing back 
along the series of Gorgonopsids to the Deinocephalia. we ought 
to have self-consistent results throughout, and to arrive at a 
condition in primitive forms which is not inconsistent with that 
found in known Reptilia. The series of Gorgonopsids at my 
disposal gives a sufticiently close morphological series to show 
with certainty that :—— 
lst. The noteh in the hinder border of the angular of a Deino- 
cephalian is homologous with the wide triangular notch 
in Diademodon. 
Ind. That the very small but distinct auditory groove described 
above in Deinocephalia is homologous with that of Dia- 
demodon. 
3rd. That the ridge to the outer side of the paroccipital on the 
squamosal in the two types is the same. 
These homologies lead us to expect that the tympanic membrane 
of Deinocephalia should be attached to the squamosal just outside 
the paroccipital process, to the posterior surface of the quadrate, 
and to the edge of the reflected lamina of the angular. I believe 
such an attachment to be workable inasmuch as a tympanic 
membrane attached to it might be flat. 
The groove for the external auditory meatus in Deinocephalia 
is of great interest. It is undoubtedly homologous with that of 
Diademodon, and I think almost as certainly with that of a 
mammal. It gives clear evidence that even at this time the 
tympanic membrane (whatever its precise situation and attach- 
ments) had sunk in and was no longer directly exposed on the 
outside of the head. It 1s formed by a groove between the outer 
edge of the tabulare and the squamosal; this groove terminates 
suddenly where the tabulare ends in a definite projection from 
the back of the skull which must mark the upper insertion of 
the membrane. 
It is known from the evidence of specimens which show the 
complete stapes in position—Hryops, Trimerorachis, Cyclotosaurus, 
ete.—that in Stegocephalia the tympanic membrane was stretched 
between the tabulare and the squamosal across the otic notch 
on the upper surface of the skull. By direct tracing back of 
homologous parts we have seen that in Deinocephalia the upper 
end of the membrane should be attached to the squamosal and 
the extreme distal end of the tabulare: that is, in the exact 
position we know it to have been attached in Stegocephalia 
related, although remotely, to the Deinocephalian’s amphibian 
ancestors. 
