GREGORY, PRESENT STATUS OF ORIGIN OF TETRAPODA 375 



The expanded iTpper portion of the cleithrum may have served for the attachment of 

 the trapezius muscle {of. Fig. 14, F). 



D, Shoulder-girdle of a Permocarhoniferous temnospondyl Cacops aspidephorus. After 

 Williston. 



The cleithrum extends over the top of the scapula after the fashion of a suprascapula 

 (cf. Fig. 14, F). The lower end of the cleithrum may represent the mammalian acro- 

 mion (Gaudry, Broom). 



E, Shoulder-girdle (ventral view) of a modern urodele, Salamandra niaciilosa, after 

 W. K. Parker. Nearly adult, enlarged 4/1. The greatly enlarged intercoracoid car- 

 tilages, which overlap in the mid-line, furnish a broad base for the pectoral muscles. 

 The anterior extension occupies the position of the epicoracoid. The interclavicle and 

 clavicle are lost, st, sternum. 



F, Shoulder-girdle (left side, outer view) of Salamandra maculosa, after W. K. Parker. 

 Nearly adult, enlarged 5/1. 



The coracoid, metacoid and scapula are still separated by sutures. The dermal ele- 

 ments are lost, except possibly the cleithrum (fcl), which may be represented by the 

 so-called suprascapula ; this gives attachment to the trapezius muscle. 



G, Shoulder-girdle of a Permocarhoniferous branchiosaur, Brachiosaarus salaman- 

 droides. After Fritsch. Viewed from above, x 6/1. In the branchiosaurs, which are 

 in many characters structurally ancestral to such modern urodeles as Crypto'branc'hus, 

 the intercoracoids are much expanded, the scapula has a short truncate blade and the 

 dermal elements (clavicle, interclavicle, cleithrum) are slender. 



H, Scapulocoracoid of a Permocarhoniferous reptile, Dimetrodon. After Williston. 

 The expansion of the coracoid and the small size of the metacoracoid are reminiscent 

 of the temnospondylous Amphibia. This scapulocoracoid is fundamentally similar to 

 that of the Therocephalia and doubtless represents a pre-mammalian stage, acr, acromial 

 process. 



In short we conclude : ( 1 ) that in primitive reptiles the homologne of 

 the true mammalian coracoid is that anterior element which: {a) lies 

 below the acromial border of the scapula, (&) above th^ clavicle, (c) 

 enters the glenoid articulation posteriorly and (d) extends ventrally 

 toward the sternum; (2) that the posterior element or metacoracoid in 

 many mammals persists in a reduced condition, arising from a separate 

 center; (3) that in monotremes, lizards, Sphenodon and certain other 

 reptiles the suture between the coracoid and metacoracoid may have 

 disappeared or else the metacoracoid has disappeared (AYilliston) ; (4) 

 that the epicoracoid is a thin membranous element lying between the 

 coracoid and the interclavicle in lizards, Sphenodon and the monotremes, 

 which was probably present also in many of the Permian Tetrapoda. 



Conclusion 



Although the foregoing discussion of the derivation of cheiropterygial 

 limbs from piscine appendages is regrettably incomplete, it is at least 

 based so far as possible upon the combined data of comparative anatomy 

 and palaeontologv, and the evidence here summarized is, I hope, in line 

 with the historical trend of development in these fields. 



From the viewpoint which has been developed in the preceding pages, 

 the origin of the Tetrapoda is regarded as a remote consequence of the 



