In the Cretaceous rocks the scapula presents, in most cases, a different appearance. The coracoid is an elongated, somewhat triangular bone, compressed on the outer margin as in birds, but differing alike from birds and other Pterodactyles in not being prolonged forward beyond the articulation for the humerus. In these Cretaceous genera, toothed and toothless alike, the articulation for the upper arm bone truncates the extremity of the coracoid, so that the bone is less like that of a bird in this feature. Perhaps it shows a modification towards the crocodilian direction. The scapula, which unites with the coracoid at about a right angle, is similarly truncated by the articular surface for the humerus; but the bone is somewhat expanded immediately beyond the articulation, and compressed; and instead of being directed backward, it is directed inward over the ribs to articulate with the neural arches of the early dorsal vertebræ in the genera found in strata associated with the Chalk. As the bone approaches this articulation, it thickens and widens a little, becoming suddenly truncated by an ovate facet, which exactly corresponds to the transversely ovate impression, concave from front to back, which is seen in the neural arches of the dorsal vertebræ on which it fits. This condition is not present in all Cretaceous Pterodactyles. It does not occur in the Kansas fossil, named by Professor Marsh, Nyctodactylus. And it appears to be absent from the Pterodactyles of the English Weald, named Ornithodesmus. [Illustration: FIG. 39. THE NOTARIUM An ossification which gives attachment to the scapulæ seen in the early dorsal vertebra of Ornithocheirus (From the Cambridge Greensand)] [Illustration: FIG. 40. RESTORATION OF THE SHOULDER-GIRDLE IN THE CRETACEOUS ORNITHOCHEIRUS Showing how the scapulæ articulate with a vertebra and the articulation of the coracoids with the sternum. The humeral articulation with the coracoid is unlike the condition shown in other Ornithosaurs] There is no approach to this transverse position of the scapulæ among birds. And while the form of the bones in the older genera of Ornithosaurs is singularly bird-like, the angular arrangement in this Cretaceous genus is obtained by closely approximating the articulations on the sternum, so that the coracoids extend outward as in reptiles, instead of forward as in birds; and the extremities of the scapulæ similarly approximate towards each other. This rather recalls the relative positions of scapula and coracoid among crocodiles. If crocodile and bird had been primitive types of animals instead of surviving types, it might almost seem as though there had been a cunning and harmonious blending of one with the other in evolving this form of shoulder-girdle. THE FORE LIMB