84 BULLETIN 110, UNITED STATES NATIONAL MUSEUM. 



present in Antrodemus and Tyrannosaurus. The right jugal measures 242 mm. 

 antero posteriorly and vertically at the center 130 mm. 



Quadrate (q.). — The quadrate in Ceratosaurus is relatively long, and in the 

 articulated skull rises much higher on the side of the head than in either Antrodemus 

 or Tyrannosaurus. Marsh says ' of the quadrate in Ceratosaurus: 



One peculiar feature of the quadrate is a stronj;; hook on the upper half of the outer aarface. Into 

 this hook a peculiar process of the quadratojugal is inserted. 



A comjiarison of these bones with those of Antrodemus shows them to be very 

 similar and that the j)ecidiar features noted by Marsh were due to post-mortem 

 distortion and not to any great structural differences. 



The so-called hook is that part of the quadrate that forms the upper boundary 

 of the foramen between the quadrate and quadratojugal. Crushing has disarticu- 

 lated the quadratojugal and forced it in beneath this hook on the quadrate, whereas 

 normally it would have been in contact with the rough sutural surface on the external 

 border of the hook, as shown in figure 2, plate IS, its inner side then forming the 

 outer boundary of the foramen, as in the Antrodemus skuU before me. 



Bauer ^ noted the presence of this foramen, but Hay' thought it absent, not 

 recognizing tlie outlines of the border of the foramen or the partial dislocation of 

 the quadratojugal. This interpretation of the relationsliips of these elements, 

 therefore, renders the lateral view of these bones in the skull as restored by Marsh 

 incorrect. The greatest length of this bone in Ceratosaurus nasicornis is 212 mm., 

 whereas in Antrodemus, representing a larger individual, it measures only 192 mm. 

 Except in minor details, other than the differences discussed above, tlie quadrate 

 closely resembles the same element in Antrodemus ralens. 



Squamosal (sq.). — Tlie squamosal is considerably weaker tlian in Antrodemus, 

 though its relations with adjacent bones appears to be almost identical. The bar 

 extending do-miward in front of the quadrate to unite with the quadratojugal is 

 exceedingly narrow anteropostcriorly, as comjiared with the same element in Antro- 

 demus. The process forming the back of the cotylus for the head of the quadrate 

 is much more slender than in the latter genus. The union of the squamosal with 

 the postorbital is not l)y a vertical suture, as indicated in the restoration of the 

 skull by Marsh (fig. 1, pi. S, Dinosaurs of North America), but is by a long squamous 

 suture (fig. 2, pi. 18). 



Palate. — The palate of the type skull of Ceratosaurus nasicornis has been described 

 by Dr. O. P. Hay,^ and with the exception of interpolated notes, his description is 

 given in its entirety: 



Figure 1 [fig. 51] represents a view of the left side of the skull, that which best shows the bones of 

 the palate. In the fossil the bones of the palate stand nearly perpendicalar, but OTthout doubt this is 

 to a great extent due to crushing. They probably sloped upward and inward at an angle of 45°, or more 

 probably they formed a vaulted roof for the mouth. 



The postero-interior process of the pterygoid (fig. 2, numeral 28) joins the inner border of the 

 quadrate at a point 55 mm. above the articulation of the latter bone vvith the lower jaw. The posterior 

 superior process rose to the upper end of the quadrate. [Evidently Hay had the idea that the pterygoid 

 extended posteriorly far enough to join the heavy vertical portion of the quadrate, whereas this bone in 



' Dinoaurs of North America, 1S36, p. 159. ' Proc. U. S. Nat. Mus., vol. 33, 1908, p. 362. 



2 American Naturalist, vol. 25, 1S91, p. 446. « Idem., pp. 363, 366. 



