32 THE CERATOPSIA 



Premaxillary 



This forms in a broad, smooth, vertical plate, the osseous septum, which divides the external 

 nares. Above it meets an equivalent portion of the nasals. The septum differs from that of 

 Triceratops in being thinner and in lacking the fenestration 4 found in all Lance genera. This vertical 

 plate is called the ? septomaxillary by Von Huene in his description of the skull of Triceratofs. 

 There is in no instance known to me any indication of suture between it and the premaxillanes proper. 

 Here, again, however, the suture may have become obliterated at an earlier age than is represented 

 by any known skull. One would expect to find it in the very youthful Brachyceratops montanensis 

 type, as well as in the skull of Triceratops flabellatus in the Yale Museum in which the cranial ele- 

 ments are practically all isolated from each other, but the latter does not show it. 



Anteriorly, the premaxillary is convex and broadened out to form the front of the nose, with 

 an ascending process reaching nearly to the base of the nasal horn. This process is overlapped later- 

 ally and posteriorly by a downwardly curved process of the nasal which together with the latter 

 forms the convex profile of the nose. The lower margin of the premaxillary is strongly convex 

 downward, as viewed laterally, in all Centrosaurus specimens, while in Triceratops this margin is 

 more nearly straight. The surface of this portion of the bone is smooth. The posterior process of 

 the premaxillary, which is included between the nasals and maxillary, is relatively deeper and less 

 slender than in Triceratops. Its form and extent, however, vary considerably. In some Centro- 

 saurus skulls it reaches the preorbital fossa (lacrymal foramen of authors), in others it is widely 

 separated from the fossa. Its contact with the lacrymal is also subject to variation although whether 

 or not this is of specific value is not clear. In the two C. fexus skulls in the American and Yale 

 museums there is such contact, but in the Ottawa one there is not, nor in the type of C. nasicornus, 

 while the skull of Centrosaurus apertus No. 4519 R.O.M., which is very similar to nasicornus, shows 

 such contact, as does skull No. 348 G.S.C. with a sharp forwardly curved nasal horn. In other 

 words, the ascending process of the premaxillary meets the lacrymal in two out of three skulls 

 identified as C. fiexus, and in one skull of nasicornus, and in one with the forwardly directed nasal 

 horn. The bones fail to meet in one nasicornus and in one fexus skull. 



Nasal 



The nasal of Centrosaurus is of very different proportions from that of Triceratops, due in part 

 to the position of the erectly arising nasal horn which, as we have seen, lies over the posterior rather 

 than the anterior margin of the nares. This makes the portion of the nasals anterior to the horn rela- 

 tively longer in the present genus. Thus, in Centrosaurus, the nasals form nearly half the anterior 

 profile of the muzzle, in Triceratops, little, if any. Posterior to the horn, the nasals of Triceratops 

 are relatively longer. On account of the depth of the face in Centrosaurus, the nasals are also 

 deeper than in Triceratops. Anterior to the horn, the nasals in Centrosaurus are divided, and are 

 separated by the anterior ascending process of the premaxillaries, which apparently is not true of 

 Triceratops. Posterior to the horn, on the other hand, the suture separating the right from the 

 left nasal is obsolete and not discernible in any skull which I have studied. In Triceratops, the 

 suture is, as a rule, clearly visible to the extreme posterior end except in very old individuals in 

 which practically all of the sutures have been obliterated. The nasals in Centrosaurus thus form the 

 top and sides of the nose to about halfway between the nasal and supraorbital horns, where they meet 

 the prefrontals above and the lacrymals (and ? preorbitals) on the side of the face. Thus the 

 bounding of the nasals is comparable to that in Triceratops, except that they never reach the pre- 

 orbital fossa. The variable extent of the posterior ascending process of the premaxillary may exclude 

 the nasal from contact with the maxillary as in the C. fexus type skull and the No. 4519 R.O.M. 

 and No. 348 G.S.C. skulls. In the C. nasicornus type, on the other hand, the contact between nasal 



4 This fenestration is not present in all Triceratofs skulls, which leads to the supposition that it may not have been 

 normally present in the living animal, but is merely an accidental post-mortem perforation through the thin septum. 



