Fossil Birds in the Marsh Collection of Yale University 63 



Gannet that, although somewhat imperfect, is so thoroughly char- 

 acteristic as to leave no doubt whatever as to the kind of bird it 

 belonged to in life. In comparing this bone with the coracoids of 

 different species of Gannets, I find it to come nearer Sula leucogaster 

 (No. 18023, Coll. U. S. Nat. Mus.) than to any other form. The 

 extinct species, however, had the coracoidal shaft somewhat stouter, 

 and the glenoidal facet for the humerus about one-third larger. With 

 respect to the imperfections (Fig. 123), there is to be noted the loss 

 in the fossil of the delicate osseous spur that springs from the lower 

 margin of the glenoid cavity, to curve forwards and outwards in the 

 direction of the summit of the bone, which is likewise broken off. 

 Finally, the infero-external angle is also fractured off and lost. 



Among the Sulidce, the sternal facet is peculiar and unicjue. On 

 the posterior aspect of the bone it is very broad, antero-posteriorly, 

 the bone being thickened above in order to accommodate it. Its 

 internal angle or termination is not carried to the mesial angle of the 

 bone. This facet is narrow and elongate on the anterior aspect of the 

 sternal extremity of this coracoid, all of which is strictly Gannet in 

 character, and can in no way be confused with any other bird. 



For this new and extinct species of the Sulidce I propose the name 

 of Sula atlantica — the Atlantic Gannet, now extinct. 



Uria antiqua Marsh. 



[Plate VIII, Fig. 56.) 



Catarractes antiqims Marsh, Amer. Joum. Sci., ser. 2, XLIX, 213. 

 Holotype. No Cat. Number. Academy Natural Sciences of Philadelphia. 

 Tarboro, Edgecomb Co., North Carolina. Miocene. 



says that "this specimen is established on a single coracoid," which he admits was 

 a very imperfect bone. He also admits that he had not a skeleton of any species 

 of Gannet at hand, wherewith to make comparisons, but that he was obh'ged to 

 rely upon a lithograph ingure of the coracoid of a fossil Sula, published by A. Milne 

 Edwards in his work on fossil birds. Cope's description is long and very elabo- 

 rate, which is usually the case when the describer is required to make a special 

 effort to convince his readers of the truth of his claims. 



I am inchned to believe, judging from Cope's figure (which is published upside 

 down) that the coracoid he discovered in Maryland never came from the skeleton 

 of a Garmet of the genus Sula. As it is much broken and chipped, little can 

 be said of its extremities; while it is quite evident that the shaft of his coracoid is 

 much too slender for a bone as it exists in any species of Sula, — that is, taking the 

 length of the bone into consideration. In any event, it is a very different coracoid 

 from the one I here figure on Plate XV (Fig. 123), as any one will appreciate who 

 will make the necessary comparisons. My figure is a reproduction of a photo- 

 graph of the bone, and I compared the specimen with the coracoids of some six or 

 seven different species of Boobies {Sula), including the Gannet {Sula bassana). 



