CHASMOSAURUS MORPHOLOGY 69 



The caudals seem to have a lighter neural arch than in thai of Monoi lonhu figured by I latcher." 



The spinous processes are fairl) tall, with a gentle decrease, as in Centrosourus. 



The shafts of the posterior cervical ribs are quite straight, more so than in Centrosourus, in 



which the first long rib, borne on presacral VIII, has a curious reversed curve. In Chasmo- 

 sjurus, the first ribs are very peculiar at the proximal end, in that the tubercular process is in line 

 with the straight shaft and is elevated high above the capitular, which in turn arises some distance 

 below, and projects at right angles to the axis of the rib. The result is that in attempting to articu- 

 late the rib, the tuberculum could not possibly meet the diapophysis of the vertebra unless the shaft 

 were thrown outward at an utterly impossible angle. The posterior ribs are slender and well arched 

 and the capitular process does not form the right angle with the shaft as in the anterior ribs but is 

 more as in Centrosourus. When articulated with the diapophyses of the vertebrae, the posterior 

 ribs are thrown outward in such a way as to give a flat back and broad abdominal cavity. Thus, the 

 mounted skeletons have a narrow chest and broad abdominal region. But the ribs are almost always 

 subject to postmortem deformation and the diapophyses in the fossil may be bent upward or down- 

 ward, thus materially altering the shape of the body from that which it possessed in life. The Yale 

 Centrosourus, as mounted, presents a very different appearance, as it is somewhat slab-sided and not 

 at all broad in the abdominal region. One cannot, however, trust too implicitly the curvature of the 

 ribs in their present condition. The last presacral ribs in Chasmosourus differ from those of Centro- 

 sourus. The posterior one curves outward and downward, passing beneath, but apparently not fusing 

 with, the ilium. The penultimate rib is much longer and heavier than the last, and curves forward 

 and outward to clear the anterior end of the ilium, according to description, although passing beneath 

 it in No. 2280 G.S.C. Thence it curves backward and slightly inward. At mid-length, this rib 

 thickens and bears a flattened rugosity on its hinder margin for attachment with the anterior end of 

 the pubis. These posterior ribs all lack the decided backward sweep seen in Centrosourus cutleri and 

 the Yale specimen of C. fle.xus. 



The only other distinction of note lies in the tarsus where there are three distal tarsalia, but this 

 may be due to preserval rather than a real distinction from Centrosaurus in which but two are known. 



One peculiarity seen both in the Toronto specimen and in one of the Ottawa animals lies in the 

 proximal end of the humerus, which has not only the large deltoid crest, but a corresponding one on 

 the opposite side of the bone. With the head of the humerus in articulation, this posterior crest 

 impinges against the ribs in such a way that the upper arm could not be swayed backward without 

 either a fracture of the ribs or a displacement of the shoulder. There seems to be something 

 mechanically wrong about either the skeletons or the way they are mounted. This difficulty did 

 not appear in posing the Yale specimen of Centrosaurus, in which the posterior crest is smaller and 

 of a different shape. It is quite evident that the coracoids are placed much too far apart in both 

 Ottawa specimens, a point which Mr. Sternberg concedes. 



INTEGUMENT 



A portion of the integument'" was preserved with the skeleton, No. 2245 G.S.C. The section, 

 measuring about \ l / 2 x 3 feet, lay over the pelvic arch and right flank, and extended from a point 

 nor far from the median line. There is no evidence of bony plates in the skin. The surface was 

 covered with "large round plates . . . arranged in irregular, longitudinal rows and . . . spaced from 

 two to four inches apart. They vary considerably in size and are not always distinctly differentiated 

 from the larger polygonal tubercles either in size or shape. The large plates, one of which is two 

 and one-fourth inches in diameter, were low, flat, circular, and are defined by a circumscribing groove. 

 The edges of many of the larger polygonal tubercles as well as the large round plates ha- 

 crinkled appearance due to short grooves placed at right angles to, and ending at the edge of the 

 plates. Closely surrounding these large plates are smaller polygonal tubercles giving the appearance 



" Hatcher, M.irsh, Lull, 1907, Fig. 81. 

 32 Lambc, L. M., 1914, A, PI. XIV. 



