4 Mr. A. S. Woodward on some 



portion here indicated. The outer margin of the bone is 

 excavated by the orbit (orb.), and exhibits a short and broad 

 rounded process, extending downwards over the cheek near 

 its anterior end, apparently covering the lateral ethmoidal 

 (prefrontal) cartilage ; posteriorly it first meets the squamosal 

 (sq.), which also enters the border of the orbit, and then it is 

 excavated by a deep re-entering angle, which divides it into 

 an outer small, and inner large, posteriorly directed process ; 

 inwardly it is separated from its fellow of the opposite side by 

 the supraoccipital (s.occ.) and another median element (p.eth.) 

 behind, by the ordinary ethmoidal plates (eth.) in front, and 

 only meets between these for the short length already 

 mentioned ; anteriorly the margin is not directly transverse, 

 but inclined much forwards as it approaches the outer edge of 

 the skull. The squamosal (sq.) is the last readily determin- 

 able element of the cranial roof, and in connexion with the 

 specimen now being described it is interesting also to notice 

 the detached examples of the same bone in the British 

 Museum fossil no. P. 6834. Nearly one third of the outer 

 margin of this bone enters the rim of the orbit, its hinder two 

 thirds forming an almost straight edge above the opercular 

 region ; the posterior margin is at right angles to the latter, 

 and the bone unites with adjoining elements in a very wavy 

 suture. It is much broader behind than in front ; and imme- 

 diately behind the orbital rim there is a rugose facette on its 

 inferior face for the articulation of the sphenotic (postfrontal). 

 The smooth hyomandibular facette, which is not much ex- 

 tended, can also be observed ; but the remainder of the 

 attached surface of the bone exhibits no facettes. Of the 

 problematical roof- bones, those between the frontal, squamosal, 

 and supraoccipital (marked x in PI. I. fig. 1 a) are the most 

 difficult to understand. On the right side of the fossil, fitting 

 into the posterior re-entering angle of the frontal, there is an 

 irregularly triangular plate not extending so far as the 

 occiput, and this element is apparently subdivided by a 

 coalesced suture which obliquely crosses its middle. Another 

 plate, not clearly exhibiting even coalesced sutures, completes 

 the filling of the space between the last-described plate, the 

 squamosal, frontal, and supraoccipital. On the left side of 

 the fossil the same area is occupied by more numerous plates. 

 One apparently equals the foremost triangular bone, but does 

 not exhibit any trace of former subdivision. At least four 

 elements correspond with the hinder bone, the two longitudinal 

 sutures shown in PI. I. fig. 1 a being open and the short 

 transverse suture near the back apparently closed. The 

 median plate (p.eth.) interposed between the posterior portion 



