1899.] OSTEOLOGY OF THE TUBINAUES. 403 



between these and that of Diomedea seem sufficiently marked to 

 entitle them to the ranlt of genera. The collection contains 

 complete skeletons of two species of Diomedea. 



The hemipterygoid of the Petrel is here described and figured 

 (PI. XXIII. figs, y, 4) for the first time. 



The indications of the Ciconiiform affinities of tlie Petrels 

 pointed out by other writers have been verified and additional 

 points brought to light. It would seem that the Petrels must be 

 regarded as a very ancient group, undoubtedly by no means 

 remotely allied to the Sphenisci, Colymbi, and the Ciconiiformes. 

 Their Ciconiiform affinities are most clearly seen perhaps through 

 the palate. That of Diomedea, for instance, presents many points 

 in common both with Fregata and with Oico7iia that can hardly 

 be attributed to any other source than that of derivation from "a 

 common ancestor. The holorhinal nai'es, the temporal fossae, and 

 deep supra-orbital grooves they share in common with the Penguins 

 and the Divers. The pelvis of the Procellariidse seems to be 

 traceable to a form most closely resembling that of the Penguins. 

 That of the Diomedeid* is more specialized, and in the adult, at 

 least, resembles not a little that of the Ciconiiform type. Besides 

 the Petrels, the Grebes and Divers are the only other birds which 

 have the cnemial crest greatly developed so as to rise high above 

 the articular surface of the femur. This can hardly be regarded 

 as an adaptation in the case of the Petrels, for they are not great 

 swimmers, and do not therefore use their legs as do the Divers. 



As to the arrangement of the group in the present paper, I can 

 only regret my inability to adopt in toto that of any of those to 

 whose works we are so greatly indebted ; it is to be hoped that in 

 the near future some sort of harmony will come of the existing 

 somewhat unsatisfactory state of affairs. The present scheme — as 

 adopted in this paper — though based largely on the osteologv, 

 is not entirely founded thereon ; but has been framed with a diie 

 regard to the claims of other anatomical facts. 



xi. Key to the Osteology oe the Tubinaees. 

 A. Skttll. (Plates XXIL, XXIII.) 



The skull is holorhinal and schizognathous ; with more or less deep supra- 

 orbital grooves ; a large, laterally expanded vomer fused posteriorly with the 

 palatines ; an olfactory cavity of great size ; a large antorbital plate ; and a 

 hooked upper jaw. 



A. Supra-orbital grooves without an external overhanging ledge ; temporal 

 fossae, when present, in the form of deep depressions approaching one 

 another in the middle line, and tending to cut off the cerebral from the 

 cerebellar portions of the skull ; extei-nal nares large, divided into right 

 and left apertures by a narrow bar of bone in the mid-dorsal line ; length 

 of the upper jaw never greatly exceeding that of the cranium ; orbito- 

 spbenoid imperfectly ossified ; basipterygoid processes well developed or in 

 the form of minute prickles; with a conspicuous tubular parasphenoidal 

 pneumatic aperture opening downwards above the Eustachian grooves; 

 palatines long, sharply defined anteriorly at their junction with the maxillo- 

 palatine processes, which are small and plate-like lamellee never projecting 



