7 
798 REPORTS ON THE STATE OF SCIENCE.—1913. 
once maintained, but that both types of marine turtles have evolved 
independently during a not inconsiderable time.'® The common 
cryptodirous ancestor was probably a littoral form, as the special type 
of the cervical vertebral column and of the nasal cavities seems to be 
an adaptation to marine life acquired before the lines of development 
separated.?° 
Now, with this view of the phylogeny of Dermochelys, how 
are we to interprete the shell of this animal? How can we solve the 
problem that in one group of the Testudinata, in the marine turtles, 
the Leatherback and its few allies could acquire so aberrant a shell? 
Dollo (1901) maintained that this shell is an entirely new forma- 
tion, the old shell being nearly lost during life in the open sea, and 
then the new mosaic shell being developed under the influence of a 
temporary littoral life. This hypothesis concurs well with the 
relationship of Dermochelys with the Chelontide.  Vélker (1913), 
however, has brought forward a somewhat different hypothesis,?? that 
seems to have some advantage over that of Dollo. Vd6lker sees in the 
shell of the Leatherback, not an entirely new formation, but the result 
of a secondary proliferation of old dermal ossifications present in its 
thecophorous ancestors, especially of the marginals. This view is based 
largely on facts and reflections published by Hay. Hay ** contended 
that the primitive Testudinata possessed a double shell, the typical 
thecophorous shell having been covered by a more superficial layer of 
bones, forming seven longitudinal zones on the carapace and five on 
the plastron. The vast majority of turtles he assumed to have lost 
the superficial (epithecal) layer, Dermochelys only retaining it and 
losing nearly the entire thecophorous shell, the deeper or thecal layer.** 
To support this view Hay brought forward evidence that epithecal 
elements, representing the shell of the Leatherback, were present in 
primitive Testudinata. This last contention is confirmed by the investi- 
gations of Newman (1906), Wieland,** Vélker, and Menger. The 
evidence may be summarised as follows :— 
1. In Tozxochelys bauri, a primitive marine turtle from the Upper 
Cretaceous, there are lying on the typical neurals in a discontinuous 
median series three (or four) small independent epithecal ossicles 
(fig. 4), called epineurals by Wieland **; epineurals were also present 
in other species of Toxochelys.?® Corresponding epithecal elements 
have been found by Newman?’ under the keels of the second, third, 
fourth and fifth neural horny shields in the living tortoise Graptemys 
geographica, and less well developed in Gr. pseudogeographica, in 
exactly the same position as the epineurals of Tozochelys. 
2. In Archelon ischyros, a marine turtle (fam. Protostegid@) from 
19 Van Bemmelen, Case, Wieland, Nick, Volker. 
20 Compare Nick, p. 207-208, and Volker, p. 511; probably the ancestors of 
Dermochelys took early to life in the open sea, whilst those of the Cheloniide clung 
more to a littoral life. 
21 Compare also Nick, p. 209-211. 
22 1898; 1905, p. 163; 1908, p. 17; compare also Wieland, 1912. 
28 Compare p. 794. 24 1905, p. 332; 1909. 25 1905, p. 325. 
26 Hay, 1908, p. 164; Case, 1898, p. 382. 27 1906, p. 104. 
