118 



THE AMERICAN NATURALIST [Vol. XLI 



I find twenty-one free vertebrae between the skull and sacrum. 

 Of these the thirteenth, fourteenth and fifteenth support a free 

 pair of ribs; they being quite rudimentary upon the first two, but 

 are long and slender on the fifteenth vertebra, and are without 

 unciform appendages. The following six vertebrae have ribs 

 that connect with the sternum by costal hsem apophysis. There 

 is also a pair of sacral or pelvic ribs, but their hsemapophyses 

 fail to reach the sternum, and their lower ends make extensive 

 articulation with the last pair of true costal ribs, at some distance 

 above the costal border of the sternum. The pelvis very much 

 resembles the pelvis of Rodger's Fulmar figured by me in the Pro- 

 ceedings of the U. S. National Museum (cited above), and there 

 are eight free caudal vertebrae plus a somew^hat elongated pygo- 

 style. 



The costal border of the sternum is characteristically wide 

 from side to side, and the pits between the six facettes, unmarked 

 by pneumatic openings, are very^ shallow\ The sternums of these 

 shearwaters agree in their general characters with those of the 

 fulmars. 



In P. horealis the xiphoidal extremity is doubly notched upon 

 either side of the sternal keel, and the form of the bone is tlicrv 

 symmetrical. This is not the case wdth the xiphoidal extrt'iiiily 

 of the sternum of my specimen of Puffinus creatopus. In it, not 

 only is the left side of the bone somewhat longer than the right, 

 but instead of showing the two usual notches of the right, it has 

 three, which appears to have been caused by a bifurcation of the 

 inner xiphoidal process. These inner xiphoidal processes in P. 

 ohscurus are wonderfully slender. 



The shoulder-girdle is much like that of Daption capensis, and 

 in Figure 1 I present those parts in that species articulated in 

 situ with the sternum. This figure originally illustrated a paper 

 of mine which appeared in the Proceedings of the U. S. National 

 Museum for 1S87 (fig. 1, v. X. p. 379), where the skull is likewise 

 described in connection with other observations npon tlie osteology 

 of the Tubinarcs, and these should be read in .■,,niircii..n witli the 



Ueturning to tlie shearwaters, I may say that the arrangement 

 of the })ones of the shoulder-girdle in some of them is as we find 



