162 ODONTORNITHES. 



processes have their faces for articulation with the ilium directed outward 

 and upward, and are supported by a shaft of greater or less length. In 

 the region of the next three vertebrae, the sacrum has been crushed, but 

 the processes appear to have been nearly sessile, and to have abutted 

 almost vertically ag-ainst the ilia. 



The sacrum of Ichtliyornis differs from that of the Tern in having 

 only ten vertebrae in the coossified sacral series, while there are thirteen 

 in Sterna regia. The sacrum appears also to have been somewhat nar- 

 rower in the dorsal region. The strong transverse bar of bone running 

 across between the ilia, just behind the acetabula, and formed in the Tern 

 by the transverse processes of the ninth sacral, is in Ichtliyornis formed by 

 the processes of the seventh vertebra. This is followed by only three 

 more vertebrae, instead of four as in Sterna. The posterior portion of the 

 sacrum is thus seen to be much shorter than in Hesperomis, and the 

 whole structure is broader, and not so perfectly coossified, indicating a 

 correspondingly less degree of strength in the pelvic region, and less use 

 of the legs. 



In the type specimen of Apatornis (number 1451), the portion of the 

 sacrum preserved is, unfortunately, crushed from a horizontal, instead of 

 a vertical direction, as in Ichtliyornis dispar, so that a comparison of the 

 two is made difficult. The anterior end is, moreover, wanting, but there 

 are clear indications of ten united vertebrae, and probably at least one 

 more in front of those preserved. The neural spines are united with 

 coossified tendons, and the latter do not appear to expand posteriorly over 

 as broad a region as in Ichtliyornis. The transverse processes preserved 

 in this region appear to correspond with the second and third of the 

 Ichtliyornis sacrum. The portion of the united central column preserved 

 presents, at its anterior end, faint indications of enlargement toward its 

 anterior articular face, which was doubtless, as in Ichtliyornis, consider- 

 ably larger than the centra just behind it. 



Of the transverse processes, the second preserved is much more 

 robust than the one in front of it, but the four following are lost. 



