In the skeleton contained within the box thus formed is the very peculiar pectoral girdle, composed of scapula and coracoid, the scapula so peculiar that the controversy as to its homologies is not yet quite settled. Most authors, until recently, have believed that its peculiar shape (Fig. 113) is due to the co-ossification of the procoracoid with the scapula instead of as usual its loss or union with the true coracoid, so called. We are now pretty sure that this is not true, since in reality there is no such bone as the procoracoid, the bone so called being the real or true coracoid; and because, in the second place, the long anterior projection called the procoracoid is really only an outgrowth of the scapula itself and not a fused, separate bone. Hence the bone is properly called the scapula-proscapula, and not the scapula procoracoid, as it usually has been. The coracoids are elongate and flattened and without the usual supracoracoid foramen, so generally present in reptiles. The only other reptiles having a similar structure of the scapula are the plesiosaurs, and it has been because of this apparent resemblance that some good paleontologists have thought the turtles and plesiosaurs were allied. The sacrum is composed of two vertebrae only, and the pelvis of the usual three bones, the ilium, the ischium, and the pubis, all covered over by the shell. [Illustration: FIG. 114.—Pelvis of _Chelone_, from below: _pu_, pubis: _is_, ischium; _il_, ilium (in acetabulum).] In every known turtle the neck is composed invariably of eight vertebrae, but they are peculiar in many respects. In the earliest known turtles the neck vertebrae were, as would be supposed, biconcave, but they soon became very variable in all; in each neck some are biconcave, some biconvex, some opisthocoelous, and some procoelous. And Dr. Hay tells us that the neck has increased in length in the later forms.