PALATOQUADBATE WITH NEUROCRANIUM IN CGELACANTHIDS. 53 



one ventral. The body of the bone is said to be triangular in 

 shape, with the point below, and from there to extend dorso- 

 posteriorly to its thickened base, which is said to lie strikingly 

 high, leaving only a narrow space betAveen it and the roof of the 

 cranium. The bone, as shown in the figures given, has a shape 

 and position which show that it must have occupied, and have 

 completely filled, the hollow of a large cephalic flexure of the 

 brain, both its anterior and its posterior surfaces sloping antero- 

 ventrally. The doi"so-posterior edge of the bone is concave, with 

 a short process on either side, and looks like the dorsum sellaa of 

 the skulls of certain vertebrates turned upward and backward. 

 The dorsal process of either side projects dorso-antero-laterally 

 from the corpus, and, in Wimania, is, as already stated, so tall 

 that its dorso-anterior corner may be in contact with the postero- 

 lateral corner of the dermosphenotic portion of the fronto- 

 dermosphenotic. The dorsal end of the process thus lies at or 

 near the level of the roof of the cranial cavity, and it is apparently 

 with the ventral surface of the outer end of the process that the 

 metapterygoid articulates. The anterior process of the bone is 

 lamellar in form with its basal portion extending upward along 

 the internal surface of the dorsal process, and it gives support on 

 the dorso-anterior edge of its basal poition to the alisphenoid of 

 Stensio's descriptions. Dorsal to these two processes the lateral 

 wall of the cranium must have been of cartilage, and this 

 cartilage and the dorsal process Taunt together have formed a 

 laterally projecting dorsal portion of the posterior wall of the 

 orbital fossa, and hence have formed a postorbital process in the 

 sense in which that term is here employed. Stensio says there is 

 no postorbital process in these fishes, but it is possible that he 

 employs the term in a somewhat different sense. The basi- 

 sphenoid is traversed, on either side, by but a single canal, which 

 traverses the anterior process and will be later considered. 



In Diplocercides the anterior portion of the basisphenoid of 

 Wimania and Axelia is I'eplaced by a sphenoid bone, which 

 extends forward through the orbit to the ethmoidal region and 

 is everywhere in contact, dorsally, with the ventral surface of 

 the dermal bone, or bones, that here form the roof of the 

 cranium. The process e of Wimania and Axelia is, in this fish, 

 apparently simply a protuberancfe on the lateral surface of the 

 hind edge of the sphenoid, at about the middle of its height. 

 From it a ridge runs upward along the hind edge of the lateral 

 surface of the dorsal half of the sphenoid, and, as it is considered 

 to be the homologue of the alisphenoid bone of Wimania and 

 Axelia, it is called the alisphenoid wulst. The postclinoid 

 wall of this fish does not project dorsally above the level of the 

 floor of the posterior portion of the cranial cavity, as it does in 

 Wimania and Axelia. 



The truncus maxillo-mandibularis trigemini is considered by 

 Stensio to have issued from the cranium, in each of the three 

 fishes here under consideration, posterior to the alisphenoid, and 



