Yol. 6g.] skeleton or obnithodesmVs latidexs. 405 



The keel of the sternum of Ornithosaurs had apparently a totally 

 different morphological origin from that of birds. Most authors 

 have thought that the anterior spine is homologous with the inter- 

 clavicle ; some aver its homology with the episternum of crocodiles 

 and the manubrium of birds. The facts certainly appear to prove 

 that it is the interclavicle, primarily of dagger-like shape, and 

 occupying, with its posterior end, the 'primordial cleft' of the 

 sternum, and that the coracoids rested directly on the dorsalsurface 

 of the spine with the scapulae arched in the primitive position 

 towards the vertebral column. Under the influence of flight, the 

 interclavicle became arched in front and gradually pushed back- 

 wards, until we find it in Ornithodesmus vertically at right angles 

 to the lateral expansions, instead of on the same plane, and thus 

 occupying the same position as the keel in birds. As the inter- 

 clavicle bent posteriorly under the pressure, the coracoids Avorked 

 their articular facets at first to look obliquely outwards and at last 

 laterally, thus rendering possible movement from in front to the 

 side, and bringing the free ends of the scapuloe into conjunction 

 with the neural spines of the vertebra. 



There are several examples of the interclavicle retracting in a 

 forward direction, until but a vestige remains — caused mainly, I 

 believe, by the action of swimming, the limbs with their backward 

 thrust stimulating the forward thrust of the ends of the coracoids, 

 until they united, thereby squeezing out, as it were, the inter- 

 clavicle. In the Ichthyosauridae the coracoids are in conjunction 

 behind, and separated in front by the interclavicle ; and in the 

 Plesiosauridae the coracoids, with the exception of an anterior 

 fissure, united through the greater part of their length. They 

 •only require the keystone of the interclavicle, which undoubtedly 

 in an earlier ancestor, at least, divided both the pre-coracoids and 

 the coracoids. In the Xothosaurida? the coracoids unite in the 

 median line, without a cleft, the vestige of the interclavicle being 

 found as a keystone at the united end of the clavicles, the ' omo- 

 sternum ' of Hulke. In birds the articulation of the coracoids on 

 the grooved antero-lateral margins of the sternum is reptilian ; while 

 in Ornithosaurs it is on the interclavicle, which is neither reptilian ■ 

 nor avian, but ornithosaurian, and a unique modification. 



The Shoulder-Girdle. 

 Scapula and Coracoid. 



In ScaphoynatJius, Dimorphodon, Pterodaclylus, and Iihampho- 

 rhi/achus the shoulder-girdle is more primitive than in Oniithodesmus, 

 Ornithostoma, or Nyctosaurus, lacking the high specialization of these 

 genera, and more or less retaining a splint-like form. Remarkable 

 variation in detailed characters is found in species of the same 

 genus, as, for example, Pt. suevicus and Pt. longic.6U.um. The main 

 difference between Oniithodesmus and Ornithostoma and Nyctosaurus 

 is the rather slender process in the two last-named described by 



