366 Annals of the Carnegie Museum, 



with respect to these two groups of tubinarine birds. The differences 

 in the external forms of the beaks are well shown in figs. 128-130 of 

 Plate XXXI of the present contribution, fig. 128 giving the beak of 

 a typical Shearwater, and figs. 129 and 130 those of Petrels of the 

 genus Mstrelata. I am of the opinion that ^. vociferans was closely 

 related to ^. caribbcca (fig. 129), as I have attempted to show upon 

 a previous page of this memoir. 



The differences in the osseous mandibles of a Petrel {Mstrelata 

 vociferans) and a Shearwater (Puffinus Iherminieri) are easily ap- 

 preciated upon comparing those parts in figs. 5 and 6 of Plate XVI. 

 All Petrels and petrel-like birds possess osseous beaks or mandibles, 

 such as we find figured in figs. 1-5, Plate XVI; in fig. 11 of Plate 

 XVIII, as well as in Cook's and the Diving Petrels. 



The tarsometatarsiis is generally long and slender in the petrel 

 forms; shorter and stouter in the Shearwaters. (See the various 

 figures of this bone on the plates.) 



The " Cahow," then, was a Petrel of the genus ^strclata; and with 

 this point settled, I can proceed to give an account of its skeleton. 



Osteology of JEsfrelata vociferans. 



The Skull (figs. 1-5, PI. XVI). — This now extinct Petrel was, in 

 life, morphologically typical of the procellaridine genus to which it 

 belongs, for the subfossil bones at hand in such abundance are ample 

 proof that it was osteologically so; consequently, the remainder of its 

 anatomy must have been in keeping with the characters so clearly in 

 evidence in its skeleton. The skull as a whole, apart from its smaller 

 size, essentially agrees in all particulars with the skull of ^. Icssoni, 

 as figured by Forbes in his Challenger Report on Petrels, cited on a 

 former page of the present work. The hyoid arches are here miss- 

 ing; but it is more than safe to predict that they, too, agreed with 

 those elements as found in other species of this genus. 



Viewed from above, we are first to note, in this skull of ^strelata 

 vociferans, the twin elevations of the parietals, with a well-marked 

 median and rounded groove between them. In some skulls this latter 

 is carried forward faintly to the " cranio-facial hinge." Generally 

 it is deepest at a midpoint between the orbits, where the superficies 

 between the supraorbital glandular depressions most nearly approach 



