OSTEOLOGY OF CARNIVOROUS DINOSAURS. 87 



If the conclusions arrived at here are correct, [it should be mentioned that the conclusions reached 

 here by Hay, coincide almost exactly with the conditions found by Osborn in a well preserved skull of 

 Tyrannosaurus] the internal narial passage was bounded inwardly almost wholly by the palatine; only 

 slightly by the vomer. In any ca.se, the vomer takes a less important part in the border of the internal 

 nostril than it does in either Sphenodon or Cyclura. In Cyclura, perhaps in lizards generally, the pala- 

 tines meet in the midline, thus shutting off the pterygoids from the vomers. In Sphenodon the ptery- 

 goids join the vomers. In the broad union of the pterygoid with the quadrate and the union of the 

 pterygoid with the vomer, Ceratosaurus appears to show closer relationship with Sphenodon than with 

 the lizards and crocodiles. In the crocodiles access to the supratemporal fossa from behind is cut off 

 above the paroccipital process, but not from below. 



To what extent the internal nasal passage was bounded exteriorly by the palatine in Ceratosaurus 

 can not be determined . Possibly a process of the bone extended forward outside of the passage against 

 the inner face of the maxilla. 



A comparison of the palate of Ceratosaurus with that of Diplodocus, as represented by Marsh ' brings 

 to light both differences and resemblances that are interesting. In describing the palate of Diplodocus 

 it will be convenient to regard it as divided into anterior, second, third, and posterior fourths. In this 

 genus, as in Ceratosaurus, the pterygoids extend far forward, to come into contact with the small vomers. 

 Along the midline they are applied to each other backward nearly to a line joining the articular ends of 

 the quadrates at the hinder end of the third quarter of the length. In the lizard Cyclura they diverge 

 about the middle of the length of the skull. In Diplodocus the postero-inferior process of the pterygoid 

 is very short on account of the forward swing of the quadrate. Thus the basipterygoid processes join 

 the pterygoids but little in front of the articular ends of the quadrates. 



The anterior position of the articular ends of the quadrates has not been acquired at the expense 

 of the posterior processes of the pterygoids alone. There seems to have occurred in Diplodocus a remark- 

 able shoving forward of structures that in Ceratosaurus lie across the second and third quarters of the 

 length of the base of the skull. On examining Marsh's figure of Diplodocus it will be observed -that 

 there are four vacuities lying in a row across the palate in the second quarter length. The two of these 

 which are nearer the midline are the internal nasal passages; those lying farther outward, right and left, 

 the postpalatine vacuities. In Ceratosaurus, as in Cyclura, the latter are behind the nasal passages, 

 being separated therefrom by the palatine bones. In Diplodocus the postpalatine vacuities seem to 

 have been crowded forward alongside and outside of the nasal openings. 



Each postpalatine vacuity is mostly inclosed by the maxilla and its processes, only a little of its 

 hinder end being inclosed by the palatine and the transverse bone. The hinder end of the nasal passage 

 lies between the forks of the small palatine. The rest of its inner border is formed by the slender vomer, 

 while it is bounded in front and outwardly by processes of the maxilla. 



Several of the measurements given in the table below can not be considered 

 altogether exact, for in the case of the Ceratosaurus skull, crushing has undoubtedly 

 altered the true dimensions, while some allowance must be made for restoration 

 and inexact articulation of the elements comprising the Antrodemus skull. 



i Dinosaurs of North America, p. 177, fig. 27. 



