OSTEOLOGY OF CARNIVOROUS DINOSAURS. 83 



This bone in Ceratosaurus is rcLitively shorter than in Antrodemus, in this 

 respect approaching Tyramiosaurus. Its greatest length anteroijosteriorly is al^out 

 68 mm.; its greatest depth from the nasal to the dental border is 72 mm. 



There is present, especially on the left side, a small maxillo-premaxillary 

 fenestra as found in both Antrodemus and Tyrannosaurus. Its position, however, is 

 more ventral than in either of those genera. Marsh failed to indicate this opening 

 in his restoration of the skull. 



Jfaxillary {mx.). — The maxillary bones of both sides, except for slight distortion 

 from transverse crushing, are almost perfectly preserved. The maxillary in Cera- 

 tosaurus differs from Antrodemns in the same way as most of the other elements 

 forming the lateral view of the skull, in being relativel}^ longer and more slender 

 in their construction. There are alveoli for 15 teeth, whereas Antrodemus has from 

 16 to 17. The width anteroposterior!}^ of the superior process of the maxillary in 

 Ceratosaurus nasicornis, whicli fills the space between the antiorbital fenestra and 

 the anterior nasal opening, is 1 05 mm. wide. In the gi'eat width of this part of the 

 maxillary Ceratosaurus resembles Tyrannosaurus. 



It can not be surely determined whether there was a second antiorbital fenestra 

 in the maxillary as in Antrodemus and Tyrannosaurus. In his restoration of the 

 skull Marsh did not indicate any. Hay comments on it as follows: 



It will be observed that there is a large uneven depression in front of and partly above the antiorbital 

 vacuity [tig. 54, numeral 14]. This is not well shown in Marsh's figure. Tliis depression probably 

 corresponds to the postnartial vacuity figured by Lambe as belonging to Creosaurus and Allosaurus 

 [Antrodemus]. There is a similarly placed vacuity in the skull of Diplodocus, whose nasal openings 

 are located far toward the rear of the skull. Inasmuch as this vacuity in the Theropoda and in Diplodocus 

 appears to be wholh' in the maxillary bone, it might be well called the maxillary vacuity. On the left 

 side of the skull of our specimen of Ceratosaurus no actual opening is seen through the bone, but on the 

 right side, near the front of the depression, 47 mm. behind the nostril, there seems to be a foramen. 



I would quite agree with the latter observation, though if it does represent 

 the second antiorbital fenestra it is very much smaller than in Antrodemus. 



Quadratojugal (qj.). — Viewed laterally the quadratojugal of Ceratosaurus 

 nasicornis at the angle formed by the junction of the vertical and horizontal branches 

 is considerably narrower than in Antrodemus. The horizontal branch also is much 

 longer than in that genus, its forward termination stopping slightly short of the 

 postorbilal bar, whereas in Antrodemus it ends more posteriorly. The inner median 

 surface forms the outer boundary of the quadrate foramen. Its other relations 

 with surrounding bones are similar to Antrodemus. The greatest length of the 

 inferior part of the quadratojugal is 124 mm.; greatest length dorsoventrally is 

 131 mm. 



Juyal (}.). — The jugal is much more slender than in Antrodemus. In the 

 restored Ceratosaurus skull, as published by Marsh, he erred in his interpretation 

 of the anterior extent of the jugal as extending forward of the laclrrymal and con- 

 tributing to the formation of the boundary of the larger antiorbital fenestra. It is 

 excluded from this fenestra by the lachrymal ami jnaxillary, as may be clearly seen 

 upon examination of tlie original specimen, (pi. 17.) Hay correctly identified its 

 forward extent.' It does not appear to be perforated by a foramen such as is 



' Proc. U. S. Nat. Mus., vol. 35, 1908, p. 360, fig. 1. 



