68 



THE CERATOPSIA 



Centrosaurus or Styracosaurus. Chasmosaurus differs further from both of these genera in having a 

 low nasal horn which may be curved backward. The supraorbital horns, on the other hand, may be 

 more prominent, although extremely variable, ranging from practically nothing (see PI. VI) to a 

 remarkable pair of Cerat ops-like horns in a specimen in the American Museum (Chasmosaurus 

 kaiseni Brown, No. 5401 A.M.N. H., see PL V, A and Fig. 38). These last arise close together, 

 flare outward, and curve strongly forward and upward. In this specimen, the nasal horn apparently 

 consists of three elements, one of which may represent the espinasal, whereas in a companion skull 

 (No. 5402 A.M.N.H.) the nasal horn is more prominent, and without sutures, while the supra- 

 orbitals are small, and trihedral in section, with a rounded outer face. All of these horns have deep 

 vascular impressions. The skulls, except for certain minor variations, are quite similar and are cer- 

 tainly generically, but not specifically, identical. At the American Museum, the skull, No. 5401, 

 with small nasal and large supraorbital horns, has been considered a male, the other, with larger 

 nasal and smaller supraorbitals, a female. This so-called male skull has supraorbitals not unlike the 

 type of Ceratops montanus from the Judith River beds, which consists of supraorbital horns only, 

 with no other diagnostic parts of the skull preserved. It is the only Chasmosaurus skull known in 

 which there is the least correspondence between the two genera, which in itself is not sufficient evi- 

 dence for their synonymy. This will be further emphasized in the discussion on genera and species 

 (see p. 94). 



Reverting to the crest once more (Fig. 30), the large fenestrae lie entirely within the central 

 element (parietal), which is reduced to a rather slender bar of bone between them and a yet more 

 slender portion between fenestra and squamosal on either side. A peculiar feature is an overlapping 

 suture in this lateral bar, about mid-length of the fenestra. This apparently is not a suture between 

 two separate cranial elements, but seems to represent the place of contact of two progressive processes 

 of the same bone, one of which developed backward laterally to enclose the opening, the other for- 

 ward from the transverse posterior bar to meet, but not fuse with, the first. 50 A large epoccipital 

 accentuates each postero-lateral angle of the crest and lies just behind the posterior limit of the 

 squamosal. The form of the posterior bar of the parietal varies. Sometimes, as in the Yale and 

 American Museum specimens, it is more or less straight across, again it may be curved so as to form 

 a distinct emargination at the midline as in the Ottawa skull, No. 2280 (see p. 93). The squamosals 

 are long, extending nearly to the posterior corner of the crest, being three or four times longer than 

 wide. The inner border is thickened and rounded and the outer slightly crenulated, bearing seven 

 to nine epoccipital bones. Between the outer and inner margins the after part of the squamosal is 

 depressed, and differs from most of the dorsal surface of the skull in possessing no vascular grooves, 

 another point of similarity with the crest of Torosaurus. There are, however, vascular impressions on 

 the surface of the parietal element in Chasmosaurus which, together with the possession of epoccipi- 

 tals, distinguish it from Torosaurus, in which the entire crest, except the anterior end, is smooth. The 

 epijugals are prominent. 



VERTEBRAL COLUMN 



The vertebral column is comparable to that of Centrosaurus except for the sudden increase in the 

 height of the neural arches and spines posterior to the crest, i.e., beyond the tenth presacral vertebra. 

 Both have twenty-one presacral vertebrae ; the sacrum has the usual ten vertebrae, in four of which 

 (the second to fifth) the parapophyses or sacral ribs coalesce distally to form the acetabular bar as in 

 Centrosaurus; behind these are four other fused centra whose transverse processes do not meet the 

 ilium. The anterior fused sacral also has free diapophyses. The spinous process of sacral I is 

 isolated, those of II to VI are fused into a continuous plate, while those of Vll to X are isolated 

 again. The sacral ribs all seem to spring from the point of union of adjacent centra, whereas in 

 Centrosaurus the last arises from the middle of its centrum. 



50 In the skull of Chasmosaurus No. 2016 Y.P.M., these processes do not meet, hence a portion of the outer margin of 

 the fenestra is formed by the squamosal (see Torosaurus gladius, p. 132). This appeared in the course of further preparation 

 of the skull, subsequent to the drawing of the figure. 



