ON THE PHYLOGENY OF THE CARAPACE, ETC. 793 
be either a new formation or a modification of the thecophorous shell ; 
it must have developed in a relatively short time, and the problem of 
its origin grows more puzzling and more interesting. We must first 
consider the evidence that the Leatherback belongs to the Cryptodirous 
Testudinata and is allied to the other marine turtles, Chelonia and 
Thalassochelys (Cheloniide), and related fossil forms (e.g. Tozrochely- 
ide and Protostegide). We must do so carefully, as the body of 
evidence considered convincing by Baur and others was judged in- 
sufficient by such competent authorities as Hay, Gadow (1901), and 
Boulenger. 
It is well known that nearly all Testudinata can retract their head 
and neck within the shell, and that they do so in two different ways. 
In one group, the Pleurodira, the neck is bent sideways; in another, 
the Cryptodira, in a vertical plane, in a somewhat S-shaped curve. If 
we now compare the mode of articulation of the centra of the eight 
cervical vertebree we find, somewhat masked by a number of aberrant 
cases, that there is a differentiation of these vertebre into two groups, 
an anterior one with opisthoccelous, and a posterior with proccelous 
centra; the connecting vertebra is biconvex. In the Cryptodira it is 
typically thus: 
1(2(3(4)5)6)7)8) 
It is obvious that this differentiation is an adaptation acquired in 
connection with the sharp bending of the neck when retracted; this is 
confirmed by its absence in those Plewrodira, which do not bend their 
neck in a double curve. 
The typical marine turtles, the Cheloniide, also show this differen- 
tiation of the cervical vertebre in two groups, and details prove that 
the cervical vertebral column is of the cryptodirous type. This cannot 
surprise us, aS anatomy and paleontology prove that they are 
descendants from Cryptodira. The neck is shortened, and they can no 
longer retract their head, or even the neck, into the shell, but this 
is easily understood as a consequence of marine life. 
The neck of the Leatherback is still shorter than in Chelonia,* and 
the centra of its vertebre are united by thick cartilaginous pads, 
strengthened by strong fibrous tissue. Obviously the Leatherback 
cannot bend its neck in a strong curve. And yet we find the same 
differentiation of the cervical vertebre into two groups as in the 
Cryptodira and in some Pleurodira. Certainly this complication cannot 
have been recently evolved in the Leatherback, with the obviously 
reduced articulations of its cervical centra; it can only be understood 
if we assume that these articulations were well developed in the 
ancestors of the Leatherback, and that these animals could bend their 
neck in a very strong curve. That this bending of the neck was 
performed in a vertical plane, as in the Cryptodira, and not sideways, 
as in the Pleurodira, is shown by the great resemblance in detail of 
the cervical vertebre of the Leatherback with those of the Cheloniide, 
whilst no traces of the adaptations peculiar to the Pleurodira are 
found. We must conclude that the ancestors of the Leatherback could 
4 Volker, 1913, p. 493. 
