40 THE CERATOPSIA 



vertebra is elliptical in shape and is relatively longer and lower as compared with its nearly circular 



form in Triceratofs. . ,. ., 



Contrary to the descriptions of Marsh and Hatcher, Lull 11 recognized a suture apparently divid- 

 ing the atlas, on the strength of which he described four coalesced vertebrae instead of three. To 

 this Brown takes exception, claiming that the four Monoclones specimens in the American Museum 

 collections, and also a specimen of Triceratofs, show only three vertebral divisions clearly marked 

 by the point of union of the vertebrae, neural arches, and spines. I am now also in agreement with 

 Marsh and Hatcher as well as with Brown. Nevertheless, the sutural marking is visible not only in 

 the Triceratofs frorsus type at Yale, but also in the Yale Centrosaurus specimen. It has been sug- 

 gested that the portion of the atlas in front of this suture may represent a proatlas, which would then 

 not increase the number of vertebrae in the coossified part beyond the generally accepted three. 



The neural canal cannot be traced at its anterior end, as it is filled with matrix practically indis- 

 tinguishable from the bone. Posteriorly, however, the outline is clear. Here the canal has a 

 triangular section, broader at the base than on either side. 



The articular face for the reception of the occipital condyle is a hemispherical cup, practically 

 circular in outline, as in Triceratofs. There is little if any flattening above or below until near the 

 point where the axis and atlas meet, where the former is flattened. In Triceratofs there is a dis- 

 tinct, median, fore and aft groove on the inferior face of the axis which is not evident, or very 

 slightly so at the anterior end, in Centrosaurus. Laterally, the lower half of the centra of the axis 

 and third cervical in Centrosaurus is constricted, much more so than in the axis of Triceratofs, cervical 

 III in each being more alike. The junction of the second and third is narrower below than in 

 Triceratofs, and, if one may judge from Hatcher's figure, 12 than in M. crassus also. There are slight 

 prominences on either side of the atlas in the Yale specimen suggestive of those in Monoclonius 

 crassus, but there is no real evidence of an atlar rib such as Hatcher assumes for the latter. These 

 prominences, or at any rate the inferior ones, are lacking in the Triceratofs frorsus type, which is 

 exceptionally well preserved. 



The posterior face of the centrum of the third cervical is concave in the Yale Centrosaurus and 

 its transverse diameter exceeds the vertical, the reverse being true in M. crassus, according to Hatcher, 

 although, as he says, the proportions have been much altered by crushing. 



Cervicals I-III of the Yale Centrosaurus as well as of a specimen of Centrosaurus afertus, 

 No. 4519 Toronto, are very different in general appearance from those of the M. crassus type, 

 especially in the greater elevation of the neural spines, the fore and aft length of the third cervical, 

 and the general proportions of the entire mass. 



Cervical IV 



As it is now preserved, the centrum is opisthocoelous with the anterior face flat, the posterior 

 one concave. The anterior aspect is almost pentagonal, but the ventral angle fades out posteriorly. 

 The lateral sides are concave. Part of the right rib is present, consisting of both capitulum and 

 tuberculum. The cervicals differ only slightly from those of Triceratofs, except for size. 



Cervicals V and VI 



Cervicals V and VI (Fig. 8), evidently pathologic, are coossified, with no visible sign of division 

 between the centra. The neural arches are, however, normally distinct. The centra have a flattened 

 lower aspect and no decided constriction on their lateral faces. The articular face of the fifth is 

 hexagonal while the posterior face of the sixth, as preserved, is nearly quadrangular. 



11 Lull, R. S., in Hatcher, Marsh, Lull, 1907, p. 47. 



12 Hatcher, Marsh, Lull, 1907, Fig. 78. 



