Yol. 69."] SKELETON OP ORNI1HODESMUS LATIDENS. 403 



The Notarium. 



It is curious that in the process of development of a notarium 

 Ornithodesmus should differ so much from Omithostoma (Pteran- 

 oclon). This is shown in the absence of the supra-neural plate 

 and of the fusing of the extremities of the transverse processes 

 by a hand-like ossification. In Omithostoma the scapular union 

 took place on the supra-neural plate, and in Ornithodesmus on the 

 fused neural spines. In the former, eight vertebrae comprise 

 this compound bone ; in the latter, six. In the former, the trans- 

 verse processes of the first three vertebrae are fused with stout 

 ribs ; in the latter this feature is not seen. In Omithostoma the 

 transverse processes are of the same length ; while in Ornithodesmus 

 the median pair are shorter than the others. The reason of this is 

 not apparent. The style-like process from the posterior side of the 

 extremities of two of these may be an incipient stage of their 

 fusion. 



It appears from the notarium of Ornithodesmus latidens that the 

 six anchylosed centra, described by H. G. Seeley as the sacrum of 

 0. cluniculus, 1 belong to the notarium. "Whether they belong to 

 the notarium or to the sacrum, they are specifically separated 

 from the former by the following characters : — the centra of 

 the first and second vertebrae are comparatively flat and broad, 

 with a pronounced longitudinal valley on the ventral surface of 

 the third to the sixth, and the last four are broader and flatter 

 than the first two. The valley is absent in 0. latidens, and the 

 ventral surfaces of all the vertebrae are convex from side to side 

 and concave longitudinally. 



The Sternum. 



In 0. latidens alone, among all known forms, is there a carina 

 for the whole length of the sternum developed so highly, so arched,, 

 and with the lateral expansions so narrowed, as to approximate very 

 closely to the similar structure in birds. It would seem that the- 

 expansion of the lateral plates decreases, as in birds, in ratio to the 

 height of the keel. In 0. latidens the position of the coracoid 

 facets differs from that seen in birds. According to Seeley, 2 the 

 articular surfaces ' obliquely overlap, practically as in wading birds 

 like the heron.' In Ardea cinerea, as in all those birds that I 

 have examined, whether in the fine series preserved in the Museum 

 of the Royal College of Surgeons or elsewhere, the articular sur- 

 faces of the coracoids are situated not on the keel, but on the 

 anterior boder of the lateral expansion, and separated in nearly all 

 forms one from the other by bone. They must necessarily be 

 oblique, for the lateral expansions are so. In the heron the extreme 

 inner angle of one coracoid decussates over the other, and 



1 Q. J. G. S. vol. xliii (1887) p. 206. 



2 'Dragons of the Air' 1901, p. 174. 



