
634 The American Naturalist. -T 
supraoccipital process is club-shaped and enormously developed, 
—more than in any other tortoise known. Of course this 
character alone would not be sufficient to establish absolutely the 
near affinity of the peculiar form with the Trionychia; such a 
development of the posterior portion of the skull could take place 
in the Pleurodira or Cryptodira just as well. But there are other 
characters which at once show that the form has nothing to do 
with the Pleurodira. Before all, the pterygoids extend behind 
between quadrate, basisphenoid, basioccipital, a condition never 
seen in the Pleurodira. Whether the pterygoids are completely 
separated by the basisphenoid as in the Trionychia cannot be 
seen from the photographs; this question, therefore, is still an 
open one. The quadrate is peculiar. The articular face with the 
lower jaw is Trionychian, not Pleurodiran ; and so is the posterior 
end of the lower jaw. The quadrate is not completely closed 
behind, but only on its outer border, as in Podocnemis, for instance, 
but not in such a great degree. As is well known, the quadrate _ 
of the Trionychia is completely closed behind ; this, of course, is 
a secondary condition, and there cannot be any doubt that the 
ancestors of the Trionychia had the quadrate open behind. The 
quadrate of Carettochelys is exactly of such a form which we 
may expect in the ancestors of the Trionychia. The pterygoids 
resemble very much the same elements in the Trionychia. The 
lewer jaw is rounded in front and has a short symphysis. The 
upper side of the skull is very interesting. The greatest peculi- 
arity is that the upper surface of the bones is granulated exactly 
as the shell. The dermal plates described by Ramsay do not exist; 
there are no plates on the skull at all. This peculiar condition 
is only found in the Jurassic Compsemys plicatulus Cope. The 
sutures of the bones of the upper side of the skull, which can be 
seen, just as the sutures of the elements of the carapace and 5 
plastron are visible, must have been taken as indications of dermal nt 
plates by Prof. Ramsay. 
The interorbital Space is very large, the orbits being com — a) 
pletely lateral ; the postorbital arch is about half of the interorbital 
space. The whole upper aspect of the skull reminds us of the 
Dermatemydidæ, Staurotypidæ, Cinosternidæ; and the arrange — 
on ella he alta, 




seme 
