1012 MR. D. M. S. WATSON ON 
arrived at by the eating away of a solid cranial roof either from 
the back or from below. ‘The former process results in the loss 
of a parieto-squamosal connection and the retention of that 
between the jugal and quadratojugal, asin Trionyx: the latter 
produces a skull like that of Chelys, in which the squamoso- 
parietal connection alone persists. Extension of either process 
results in the complete loss of both connections, as in Cistaudo. 
‘The only view which really fits the case is that the primitive 
Chelonian skuil was completely roofed like that of Chelone or 
a Cotylosaurian reptile. This view is supported by the pre- 
ponderance of roofed skulls in early fossil forms. 
It is not at all difficult to derive the Chelonian skull from that of 
a Cotylosaurian reptile. The chief differences in the palate, the 
total loss of teeth, the reduced pterygoids and their firm union 
with the basis cranii, and the rudimentary secondary palate, are 
apparently dependent on the gradual development, of the horny 
beak. The unrelated group of the Anomodontia (Dicynodontia) 
show the gradual extension of a horny beak, followed by the 
total loss of teeth, the reduction of the pterygoid and its close 
union with the basisphenoid, and the development of a small 
secondary palate which in such features as the fusion of the 
prevomeys is extremely like that of a Chelonian. 
Many of the peculiar features of the posterior part of the 
Chelonian skull are due to the necessity of adequately supporting 
the very powerful vertically placed quadrate. 
Other remarkable features in which all Chelonians agree are the 
loss of the postfrontal and prefrontal (or lachrymal), and the 
usual absence of nasals ; with the exception of the last, these losses. 
have repeatedly occurred amongst reptiles. A much more 
striking feature is the loss of the facial processes of the pre- 
maxillee, so that the bony external nares areunited. ‘This change 
is only paralleled in the Mammalia, of which group it is one of 
the most striking peculiarities. 
‘“‘ Archichelone,” then, had a roofed skull with a primitive 
reptilian palate, teeth, probably only a squamosal of the temporal 
elements [and no facial processes of the premaxille]. 
WVeck.—Cope pointed out that all tortoises could be divided 
into two classes according to the mode of retracting the neck, 
either vertically in the Cryptodeira (to which may be added the 
Trionycids and Dermochelys) or laterally in the Pleurodeira. This 
feature to all appearances affords a foundation for an absolute 
division of all known types, a fact the meaning of which is plain. 
‘* Archichelone” must have had a long and flexible neck, capable 
of bending in all directions ; and after the development of the 
shell, w hen it was desirable to withdraw the head for protection, 
the two lines gradually became distinct, any intermediate con- 
ditions being obviously mechanically undesirable. 
, aeAsvall known Chelonia have eight cervical vertebra we are 
eae in claiming that number for “ Archichelone.” 
discussion of the trunk I shall aigsoaen 
