ihe Shoulder-girdle of a Plesiosaur. 345 



forward beneath the same bone, botli tending to increase the 

 firmness of tlieir union by addinjz; to the surface in which 

 they are in contact. The continued extension inwards of the 

 ventral bars of the scapuhii finally brings about their union in 

 the miildlc line, and at tlie same time their extension forwards 

 beneath the clavicles gradually shuts these bones off from the 

 ventral surface, so that we are met by the anomalous condi- 

 tion of the membrane-bones lying on the inner side of bones 

 developed from cartilage. That this condition is a purely 

 secondary one is clear from the foregoing description, so that 

 the argument that the elements in question cannot be true 

 clavicles on account of their deep-seated position falls to the 

 ground. 



It is evident that wlien the ventral rami of the scapulaj 

 have once met in the middle line, they usurp the function of 

 the clavicular arch as anterior ventral support of the pectoral 

 girdle to a gradually increasing extent ; this may account for 

 the reduction and great variability of the clavicles in some 

 forms (e. g. Miircmosnurvsj in which an interclavicle is 

 present). For the same reason the eventual disappearance of 

 the clavicular arch is very probable, and appears to have 

 taken place in the Cretaceous geims Elasmosaiirus. 



The changes undergone by the coracoids during growth 

 also tend to increase the rigidity of the girdle. Their median 

 symphysis is gradually prolonged both backwards and 

 forwards ; in the latter direction it is carried far in advance of 

 the glenoid cavity, and finally becomes continuous with the 

 sca})ular symphysis, so that in the adult the scapulaj and 

 coracoids form a continuous union in the mid-ventral line 

 from one end of the pectoral girdle to the other. 



How far the above explanation of the peculiarities of the 

 Elasmosaurian pectoral girdle applies to other types of struc- 

 ture, such as those occurring in PUosaurus and PlesiosauruSj 

 catmot now be considered ; but there apj)cars to be no reason 

 why they should not be regarded merely as less specialized 

 conditions of the same kind of modification. 



As to the homology of the ventral ramus of the scapula, 

 there seems no reason whatever for considering it to represent 

 a precoracoidal element fused to the scapula, all the facts 

 leading to the conclusion that it is merely a secondary out- 

 growtii from the latter. That this outgrowth is homologous 

 with the ventral ray of the Chelonian scapula is highly 

 probable ; and if it be so regarded, it is interesting to note that 

 while in the Plcsiosaurs it lies ventral to the clavicular arch, 

 in the Chelonia it is dorsal to it [i. e. to the entoplastron and 



