7 60 DR R. BROOM ON 



anterior of the two coracoidal elements in Dicynodon and other allied reptiles ; and that 

 this conclusion is correct can hardly be doubted. What has given rise, however, to 

 much debate is whether this is the homologue of the precoracoid of the Amphibia, 

 or a new element secondarily developed. It is developed as a large and independent 

 element in probably the whole of the Theromorpha, but the condition is best known in 

 Dicynodon, Pareiasaurus, and Procolophon. 



The tracing of the coracoidal elements back to the condition found in the Amphibia 

 is rendered especially difficult owing to the fact that in the Labyrinthodontia the 

 shoulder girdle was very imperfectly ossified. If, however, the girdles of modern 

 amphibia be examined it will be found that in practically every case there is a division 

 of the ventral portion into an anterior and posterior part — the anterior of which has 

 generally been referred to as the precoracoid. Among the Urodela this division of the 

 coracoidal region is most marked in Proteus and Menobranchus, while in Salamandra, 

 though the differentiation of the ventral cartilage into its two elements is less marked, 

 there are, according to Parker (4), distinct precoracoid and coracoid ossifications. 



In the Anura a distinct precoracoid and coracoid are almost always present, though 

 the ventral portion of the arch is here relatively much smaller than in the Urodela. 

 The coracoid and scapula are invariably well ossified by endostoses, but the precoracoid 

 bar as a rule remains largely cartilaginous, though supported in front by an ectostosis. 

 Whether this bony splint is, as believed by Parker, a true precoracoid ossification, or, 

 as held by the majority of comparative anatomists, including Gegenbaur (16), 

 Wiedersheim (17), and Howes (10), a true clavicle, is a point which is difficult to 

 decide. Of the two views that of Parker seems to me the more probable. That the 

 clavicular plates of the Labyrinthodonts which remain quite distinct in most reptiles, 

 should in the much more nearly related frogs be represented by bony splints intimately 

 connected with the precoracoid bars seems to me unlikely ; and, moreover, the fact that 

 in the Anura other elements usually ossified by endostoses are found to be ossified by 

 ectostoses would seem to confirm the opinion that the split developed on the surface of 

 the precoracoid cartilage is really part of the precoracoid. The difference in the mode 

 of ossification of the two coracoidal elements might have been brought about through 

 the precoracoid, owing to its being protected by the superficial dermal jalates remaining 

 in a cartilaginous state long after the coracoid had become ossified. 



Apart, however, from any difference of opinion as to whether a clavicle is present in 

 the Anura, it will be seen that in all the existing amphibia with limbs the ventral part 

 of the shoulder girdle is made up of a distinct precoracoid and coracoid. When the 

 most primitive known reptile is examined it is found that the ventral part of the 

 shoulder arch is here also divided into two elements, and I fail to see any valid reason 

 why the two elements in such an amphibian-like form as Pareiasaurus are not to be 

 regarded as the homologues of the two ventral elements in the living amphibians. 



The only argument, so far as I am aware, that has been urged against the homology 

 of the amphibian precoracoid with the anterior coracoid element in the Theromorpha 



