NO. lOGO. OSTEOLOGY Oh' VAMPTO^AURUS—GILMORE. 235 



libs. This interpretation excludes the anterior and posterior verte- 

 \)rx, which may he regarded as sacro-dorsal and sacro-caudal, re- 

 spectively, thus reducing tlie number of sacrals to five, as originally 

 determined by IMarsh for ('. dispar." Typically, there are five verte- 

 brae joined by suture, but, as shown in Plate 18, a^ the anterior one 

 Avould be considered a dorsal. Hence in ('. dhpar there are only four 

 true sacrals. 



Sacrals two and three were found to be firmly coossified (see fig. 

 17, S.^ and /S'.,), and in this res])ect quite at variance with Professor 

 Marsh's earlier determinations. In describing the type-specimen of 

 G. dispar, he says : 



This genus agrees with Laosairnis in one important character, namely, the 

 sacral vertebra? are not coossified. That this is not merely a character of im- 

 maturity is shown by some of the other vertebne in the type-specimen, which 

 have their neural arches so completely united to the centra that the suture is 

 nearly or quite obliterated. T(» this character of the sacral vertebrse, the name 

 of the present genus refers. 



While the neural arches of the specimen here considered are at- 

 tached to the centra throughout the column, the, sutures in all in- 

 stances are plainly discernible. Inasmuch as a second specimen in 

 the National Museum, Xo. 4753, the holotj^pe of C. depr^essus, has all 

 of the vertebrae of the sacral region firmly coossified, it would appear 

 that what Marsh considered a very important character of the genus, 

 namely, the noncoalescence of the sacral vertebra?, can not be relied 

 on as being a constant character. Moreover, the union of the other 

 centra in the sacrum of specimen Xo. 4282 were very close and par- 

 ticularly strong, and it was with some difficulty that they w^ere sep- 

 arated for the purpose of study. So firmly are sacrals two and three 

 coossified that in places the suture is entirelj^ obliterated (see fig. 17). 



The first sacral may be distinguished by the great transverse ex- 

 pansion of the anterior end of the centrum. Both extremities have 

 roughened sutural surfaces, which unite closely and strongly with 

 the centra both preceding and following. The anterior face is some- 

 what angularly convex, while the posterior is slightly concave. The 

 inferior surface lacks the decided keel of the last or sacro-dorsal and 

 is more evenly rounded. The neural canal is much expanded, as 

 shown in fig. 17, S^. The posterior parts of the pedicles of the neural 

 arch are very thin transversely and comparatively short antero- 

 posteriorly, being to a limited extent borne on the last dorsal. The 

 spinous process, as shown in Plate 13, is very similar to that of the 

 last dorsal described above, but is more anteriorly placed in relation 

 to the centrum. A long diapophysis extends out over the top of the 

 first sacral rib, which, in some individuals, is firmly ankylosed to 

 that bone. In No. 1877r?, however, the suture remains distinct. 



"Amer. Journ. Sci., XLVII, 1894, p. 246. 



