26G OSTEOLOGY OF TUBINARES AND STEGANOPODES. 



The coracoidal grooves are long and shallow, meeting rnesiad at the 

 raaimbrial base and having a lip of bone at their externo-iuferior bor- 

 ders to hold each coracoid in position when articulated. 



As in the sternum just described, the elements of the pectoral arch or 

 shoulder-girdle are non-pneumatic, and, in consequence, proportionately 

 heavy. 



The furcula when articulated differs from that of the Petrel in not 

 reaching the anterior border of the sternum. Viewed from in front we 

 find it to hi' of the U-shaped style, with the limbs of nearly uniform 

 caliber throughout. 



in articulation the pointed clavicular heads merely rest against the 

 inner aspect of each coracoidal summit, while their tips overlap the 

 anteromedian angle of either scapular head. Thus the tendinal canal 

 is completely closed in among the three bones. 



This furcula has no hypocleidium, but the arch is considerably thick- 

 ened at its usual site. 



A coracoid has a large tuberous head, which is bent forward and to- 

 ward the median line. The scapular process is very extensive, being 

 carried well down upon the antero posteriorly compressed shaft. I be- 

 lieve it will always be found to be pierced by the foramen. The cora- 

 coid of this Fulmar acquires a very unique form from the extraordinary 

 manner in which the infero lateral angle of its sternal extremity is pro- 

 duced. This is even more striking than we found it in the Petrel. 



The head of a scapula is broad transversely and somewhat compressed 

 from above downward. It offers about the usual amount of articular 

 surface for the glenoid cavity, and when in situ its anterior border occu- 

 pies the entire superior line of the scapular process of the corresponding 

 coracoid. 



The blade of the bone is comparatively short and gently arches over 

 the ribs in the usual manner. Its anterior two-thirds is narrow and 

 thickened, while its hinder extremity is slightly dilated and its tip 

 rounded off. 



These characters of the shoulder-girdle, as I have given them, agree 

 in the five or six specimens before me, and I have intentionally omitted 

 any slight deviation due to individual peculiarity. 



Of the pelvis and caudal vertebra 1 . — As already mentioned above, the 

 anterior vertebra of the pelvic sacrum extends beyond the iliac bones 

 (Figs. 10 and 11), and its neural spine is indistinguishably anchylosed 

 with that of the next one behind it, and so on to a point opposite the 

 acetabuhe, where this neural crest is suppressed, and the rim that sur- 

 mounts it for its entire length merges into the flattened neural arches 

 of the next three or four vertelme. Both at this point, and still more 

 so behind, these sacral vertelme are unusually well individualized, so 

 that the skeleton of the tail seems to really begin between the cotyloid 

 cavities. Usually, however, eleven or twelve are anchylosed in the 

 "sacrum" and eight or nine are free and constitute the tail, in addition 

 to the terminal pygostyle. 



