SYSTEMATIC POSITION OF THE PETRELS. 



In all these characters the Diomedeidce differ from the rest of the group. Further, 

 this family must be divided into three genera — Diomedea, Thalassogeron and Phcebetria, 

 of which Phcebetria is probably the most primitive. 



The evidences of the Ciconiiform affinities of the Petrels pointed out by 

 other writers, are quite borne out by a study of their osteology which, furthermore, 

 supports the contention that we may claim affinities of the Petrels both with 

 the Divers and the Penguins. The last two, it may be remarked, like the 

 Petrels, have schizognathous palates and holorhinal nares and deep supra-orbital 

 grooves. It may well be that these three, with the Ciconii formes, represent so 

 many branches of a common stock. The fact that, on the whole, the Petrels, 

 osteologically, approach the Steganopodes is quite in accordance with the ancient 

 character of this group. But the Tubinares are apparently older than the 

 Ciconiiformes ; at any rate I find that the head of the embryo Pelican bears 

 a most striking likeness to that of a Petrel ! And this both in the form of the 

 Rhamphotheca and of the cranial characters. 



Osteology lends no support to the contention that the Petrels are related to the 

 Gulls, a contention now generally abandoned, though still finding champions. The 

 likeness between the two groups is one of homoplasy only. That is to say, the likeness 

 is superficial and wrought by the demands of a like physical environment. Though 

 leading a precisely similar life the Gulls have no hypapophyses in the thoracic 

 vertebrae, for instance, while, further, the Gulls are schizorhinal and schizognathous. 

 The Petrels, like the Penguins and Divers, are holorhinal and schizognathous. 

 Moreover the schizognathism of the Petrels differs in kind from that of the Gulls. 



Myological Characters. 



It would be foreign to the purpose of this summary to pass in review the general 

 myological characters of the Petrels. Those which seem to serve most decisively 

 as signs useful for taxonomic purposes are to be gleaned from the muscles of the 

 wings and thighs. 



The " Great Pectoral " is divisible into two layers, by an interposed tendon as in 

 the Steganopodes and Storks, while in the presence of what Forbes calls a pectoralis 

 tertius, and Gadow the pars sterno-coraco-clavicularis of the supracoracoideus, the 

 Petrels are peculiar. An expansor secundariorum is found only in Oceanites and its 

 immediate allies. As in the Gulls, the tensores patagii are complicated ; but 

 the complexity seen in the Petrels is unique in character. In a considerable 

 number of Petrels the brevis tendon develops a peculiar ossicle, met with again among 

 the Steganopodes, and curiously enough in Merops ! Into the nature of this ossicle, 

 its form, and relations, it is unnecessary to enter here ; but admirable figures 

 illustrating these points will be found in Forbes's " Memoir of the Petrels " 



XIX 



