PHYLOGENY OF THE PAL^OGNATH.E AND NEOGNATH.^. 225 



in another it was perforated by an oval foramen, which Parker suggests indicates a 

 paired origin. 



No traces of a keel were observed, though such are undoubtedly present in some 

 adults (fig. 4 d). 



Ossification of the sternum, according to Gadow, begins during the last week of 

 embryonic life. 



In Struthio and Apteryx a pair of pleurostea only are represented ; in Rhea, Casuarius, 

 and Dro7nmis there are a pair of additional centres representing pro-ostea. 



Although there is no positive trace of a keel in the developing " Hatite " sternum, it 

 is generally assumed that such a structure was present in more remote ancestors, and 

 that its loss is the result of degeneration accompanied by decaying, and terminating in 

 absolute loss of the power of flight. It is possible, however, that the earliest birds did 

 not possess a carinate sternum. The Ratite sterna of the Palwognatlid' may well 

 represent this ancient type. Tlie keel was never largely developed in the sterna of the 

 Pterodactyle, and is not always present among the Chiroptera. 



b. Pectoral Arch. 



In Struthio and Apteryx only is there a well-developed |)re-coracoid. 



Apteryx, however, presents, as Parker [72] points out, " the remarkable circumstance 

 that in one species oi Apteryx {A. oweni) the coracoid is solid, presenting no coracoid 

 fenestra and therefore no procoracoid, while in another species {A. australis) the 

 prncoracoid is present at a comparatively early age, and is frequently retained in the 

 form of a ligament in the adult. In other words, the differences between the shoulder- 

 girdle in two species of Apteryx are of precisely the same nature as those distinguishing 

 the Stritthioiies from the other Ratitse." 



In Dro7nceus and Casuarius and in Phea the pre-coracoid is vestigial. It may even 

 be, as Lindsay has suggested, a new growth of the coracoid rather than the homlogue 

 of the pre-coracoid of Struthio. 



Clavicles occur only in the adult Promceus and embryo Casuarius. 



Pelvic Arch. 



The pelvic arch of the Palceognathce differs from that of the XcognatJue chiefly in 

 two points : — (1) The great heiglit of the neural spines of the synsacrum (PL XLIV. 

 fig. 5) ; and (2) the persistent ilio-ischiadic fissure. But though there can be no 

 difficulty whatever in distinguishing the Struthious from the Neoguathine pelvis, yet 

 these cannot apparently be defined in mutually exclusive terms. There is no com- 

 prehensive definition that will cut off the one from the other. To keep to the pelvic 

 arch of the Struthious birds : it is difficult to find a definition for this which 

 shall not be subject to any exceptions ; for although the persistent ischiadic fissure is 



