So very different is the structure of its shell that some excellent naturalists regard _Dermochelys_ as the equivalent in rank of all other turtles combined, the sole representative of the suborder Athecae, as distinguished from the Thecophora. Dr. Hay, whose authority on fossil turtles is of the highest, believes that its line of ancestry has been distinct from that of all other turtles from Triassic times at least. Others believe that the leather-back is merely a highly specialized form derived from the ordinary shelled type, a descendant of some of the marine turtles of Cretaceous times. In support of the primitive ancestry of the leather-back Dr. Hay offers the following: “The writer holds the view that the earliest turtles possessed practically two kinds of shell, one purely dermal, consisting probably of a mosaic of small bones arranged in at least twelve longitudinal zones. Each zone probably consisted of a row of larger bones bordered on each side by smaller bones. Each of these bones was covered by a horny scute. The nearest approach to such a dermal shell is in our days seen in _Dermochelys_. Beneath the skin there seems to have existed a carapace more or less complete, which consisted of a nuchal, a median row of neurals, eight pairs of costals, a pygal, probably one or more supraneurals, and about eleven peripherals on each side. To what extent the neurals and the costal plates had become anchylosed to the neural spines and the ribs respectively, it is now impossible to determine. Nor can we say to what extent the various elements of the carapace had become connected with one another. There was a subdermal plastron which was composed of at least eleven bones. “According to the author’s view, as time went on the external, mosaic-like shell disappeared in most turtles, while a more efficient armor was developed out of the subdermal elements. In the ancestors of _Dermochelys_, however, the dermal armor was retained, while the more deeply seated one disappeared, with the exception of the nuchal bone.”