792 REPORTS ON THE STATE OF SCIENCE.—1913. 
are solidly united to the vertebree and the ribs. In the Leatherback 
the carapace (fig. 1) is composed of a very large number of small, thin 
bony plates, fitting together with their jagged edges, forming a mosaic; 
and this carapace is separated from the vertebre and the ribs by a 
very strong layer of connective tissue, the much thickened cutis. Of 
the whole normal carapace we recognise only one element in Dermo- 
chelys, the nuchal; it is separated from the overlying mosaic shell 
by the cutis. 
The ventral side of the trunk is covered in typical Testudinata by 
an armour of bony plates, mostly nine, of the same type as the plates 
of the carapace, and forming the plastron. These elements are very 
old, having developed from the membrane bones of the shoulder-girdle 
and from the abdominal ribs (Gastralia), perhaps strengthened by a 
layer of dermal bone.t On the ventral side of the trunk of the Leather- 
back we find a very thick cutis, and imbedded on its inner side the 
elements of the normal plastron reduced to slender bars of bone. In 
Psephophorus, an extinct tertiary ally of the Leatherback, we find a 
complete ventral armour of small dermal ossicles at the sides of the 
animal, continuous with the dorsal mosaic of the carapace. This ventral 
armour is represented in the Leatherback by a small number of 
irregular ossicles lying in a median (partly paired) and two paired 
very incomplete lateral rows, five rows in all.? This remarkable 
difference in the dermal skeleton of Dermochelys and a few fossil 
allies on one side, and all the other turtles and the tortoises on the 
other, is the more curious, as the shell of the latter shows, on the 
whole, very little variation, the plates being fixed in number within 
very narrow limits, and eventual variation is mostly in the direction 
of reduction and gives us but little help in the interpretation of the 
aberrant shell of Dermochelys. 
The problem of the phylogeny of the shell of Dermochelys would 
seem less puzzling if the Leatherback could be shown to represent a 
very primitive side-branch of the Testudinate-stem—viz. a suborder 
Atheca, all the other members of the order being united in the suborder 
Thecophora. This would allow the interpretation of the shell of 
Dermochelys as a primitive one, as a forerunner of the typical theco- 
phorous shell. 
This view, however, though adhered to by several competent 
authorities, seems untenable. It was refuted by Baur, Case, Dollo, 
van Bemmelen, and Wieland, who brought evidence that the Atheca 
are Cryptodirous Testudinata, rather closely related to the other 
marine turtles, and cannot represent a primitive offshoot, a primary 
division (sub-order) of the Testudinata. This conclusion is confirmed 
by Nick, Volker, and Menger, who have studied the Leatherback and 
the ontogeny of the shell in the Testudinata under the direction of 
Professor Spengel and myself in the Zoological Institute of the 
University of Giessen. The mosaic shell of the Atheca must, then, 
1 Cf. Menger. 2 Volker, 1913, pl. 30, fig. 2. 
8 This paper is largely a summary of the more extensive original papers by Nick 
(1912), Volker (1913), and Menger (not yet published); for all details I must refer 
to these papers. I am also greatly indebted to Professor Spengel for many useful 
suggestions. 
