96 BULLETIig- 110, UNITED STATES NATIONAL MUSEUM. 



Compared with Antrodemns, the neural spines of the cervical vertebrae are 

 relatively longer, especially in the posterior half of the series, though they go through 

 much the same change of fore and aft dimensions; that is, narrow to wide, be- 

 coming narroM' again in the posterior members. 



The sixth cervical sho^vn in plate 20, figure 5, when compared with the corre- 

 sponding vertebra of Antrodemus, shows striking differences in the structure of 

 the neural arc]i, especially in the shape and downward development of the diapo- 

 physes, the articular extremity reaching nearly to the lower median border of the 

 centrum, whereas in Antrodemus tliis end is but little below the level of the neuo- 

 central suture. This feature has perhaps been somewhat exaggerated by crushing. 

 Though none of the cervical ribs are present in Ceratosaurus the closeness of the 

 diapophysial and parapophysial facets indicates quite a different style of cervical 

 rib than is found in Antrodemus. The anterior and posterior zygapophyses also 

 occupy a lower position on the neural arch in relation to the centrum, especially 



the former. There are well-developed spurlike processes 

 extending upward and backward from the tops of the 

 postzygapophyses of the cervicals, as in Antrodemus, but 

 these are not shown in the illustrations of the cervicals on 

 plate 20, which were prepared under the direction of Marsh. 

 Since they were drawn from the left side, which is now 

 hidden from view in the mounted skeleton, it is presumed 

 they were missing from that side. The cervicals posterior 

 to the sixth have suffered to such an extent, either by crush- 

 ing or loss of parts, as to render them of little value for 

 comparison or descriptive purposes. 

 Fig. 56. Twelfth presace.u, oe Dorsol vertelrae. — In all of the thoracic region there are 

 THraDDOESAL VERTEBRA OR cer- p^rts prBscnt of 12 dorsal vertebrae in the type specimen. 



ATOSAURUS NASICOEOTS MARSH. \ ^ . , 1 -, , 1r>o\ 



No.4735,u.s.N.M.TYPE. J NAT. AH exccpt thc last of the series (dorsal 14 or presacral 23) 

 SIZE. VIEWED FROM LEFT SIDE, j^^vc mucli of thclr proccssBS missing, and a comparison of 



This drawing is deficient in • j- .li x i i i i ^ i or,\ -iU 



SO MUCH AS THE TRANSVERSE thc outlmc drawing 01 the mounted skeleton (pi. dOj witn 



process is NOT INDICATED, ^j^g gpg|^ skclctal rcstoration of this animal by Marsh (pi. 



27, upper figure), indicates that some portions of these 



bones have been lost during the considerable interval of time that has elapsed 



since Marsh's restoration was made. This observation apphes particularly to the 



vertebrae of the anterior dorsal region. 



While there is reason to beheve thenumber of presacral vertebrae in Ceratosaurus 

 to be 23, as in Antrodemus, Tyrannosaurus, and Gorgosauru^, the determination 

 of this point must await the discovery of a more perfect specimen. 



The dorsal centra, according to Marsh, have the articular ends moderately 

 biconcave. The sides and lower surfaces are concave, but are not so deeply 

 excavated as in Antrodemus, as may be seen by comparing the posterior dorsals 

 shown in plate 30, with text figirres 25, 26, and 27. The centra are relatively 

 longer than in Antrodemus and none exhibit the expanded flangelike ends of that 

 genus. 



In the anterior dorsals the centra are short, the pleuropophyses is on the side 

 of the centrum as shown in figure 56, and on the fom"teenth presacral (fifth dorsal) 



