THE SKULL 35 



variable in occurrence as well as in form. Generally, there are two, but the type of Centrosaurus 

 ifcwus has but one, that on the left side. In Lambe's type of Centrosaurus apertus, the right process 

 only is preserved. This was detached at the time of Lambe's first description and, as it was an 

 unknown element, he very naturally mistook it for a nasal horn and so described it. Later, it was 

 fitted into place. The Yale specimen has the normal two anterior processes. These arise from the 

 rear bar of the crest on its dorsal aspect and project forward and slightly downward over the fcnes- 

 trae. They taper to a point, are somewhat curved, and are irregularly grooved or sulcated on their 

 upper surface. Where the right process is missing in the type, the point where it would normally 

 have arisen is smooth and shows no evidence of former attachment. Brown speaks of these 

 processes as being of the nature of the ossified tendons which occur in this group of dinosaurs. The 

 fractured end of these processes does not seem to bear this out. Compared with sections of the 

 nasal horn core, the structure revealed is essentially the same, cancellous bone within and a denser 

 outer layer. In the nasal horn are occasional larger canals which doubtless contained bloodvessels, 

 for the vascular grooves on the outer surface begin abruptly proximally, branching distally as though 

 the bloodvessels carrying nourishment to the horny sheath penetrated from within the bone outward, 

 there to run their course in the external grooves. These are not apparent in the processes, but this is 

 the only histological difference which I can discern. On the other hand, ossified tendons from the 

 dorsal region of Triceratops brevicornus type (No. 1834 Y.P.M.) are of a very different character, 

 far denser and with a more or less fibrous outer surface. Brown also suggests that the forwardly 

 directed processes may have been for muscular attachment. If so, I cannot see what muscles these 

 could have been, especially as the ends of the processes point downward toward the two large fenes- 

 trae through which no known muscle could have passed. Nor is their direction right for a possible 

 prolongation of the temporal muscles, which have been shown to have their origin on the upper sur- 

 face of the crest just behind the supratemporal openings and, passing through the latter, to go down- 

 ward and forward within the jugals to be inserted into the coronoid process of the jaws. 



These processes constitute a sub-generic character, although as we have seen, their degree of 

 development varies with the individual, and the forward processes may be asymmetrical to the 

 extent of total absence of one, or possibly both. 



Palatal Aspect 



The Yale skull (Fig. 5), while admirably preserved on the whole, is somewhat obscure in the 

 region of the occiput, especially in the lack of clarity of the sutural limitation of the various elements. 

 The narrowness of the Centrosaurus skull is especially marked in contrast to its contemporary, 

 Chasmosaurus. Anteriorly, the rostral is well preserved, together with the maxillaries and their 

 contained teeth, the last somewhat of a rarity especially in the various skulls of Triceratops known to 

 science. The precise limitation of the individual teeth is at present obscure, so that the number may 

 not be told with exactness. There seem to be about 31 vertical rows, as compared with 35 in the 

 type of C. fiexus. 



The teeth of Centrosaurus are described by Brown as differing from those of Triceratops in that 

 the central carina of the enameled crown in the middle rows is relatively higher and the outer curva- 

 ture of the rows consequently greater. 



The rostral, seen from below, is V-shaped, and the sharp, cutting edge is prolonged backward by 

 the downwardly curved margins of the premaxillaries, although the actual edge is here somewhat 

 rounded. No trace of the rod-like vomers is present, although, as Brown says, they must have been 

 similar to those of Triceratops. These elements are so extremely fragile that they are preserved in 

 but two ceratopsian skulls so far as I know. One case is that of Triceratops serratus No. 970 in the 

 American Museum of Natural History, and the other a skull of Chasmosaurus at Yale (see Fig. 29). 

 In the C. flexus type the maxillaries are notched for articulation with the anterior ends of the missing 

 vomers. Were they present, the latter would form "a long narrow bridge dividing the space between 



•Lull, R. S., 1903, PI. LIX, and Hatcher, Marsh, Lull, 1907, Fig. 26. 



