ON CERTAIN CROSSOPTERYGIANS. 1243 



We can distinguish on the sphenoid an unpaired p)ars basalis, 

 which forms the floor of the cranial cavity, and a paired pnrs 

 lateralis, issuing upwards from the pars basalis along either 

 of its dorso-latei'al edges and extending upwards to the dermal 

 cranial roof, forming the lateral wall of the cranial cavity. 



The pars basalis coiTesponds in its whole extension to about 

 half the height of the sphenoid, its absolute height thus being- 

 much longer at the posterior end than at the anterior one. From 

 the anterior end it tirst diminishes in width backwards, so that 

 some distance anterior to and below the opticus exit (II) it 

 forms only a rather thin wall between the orbits (text-fig. 2). 

 Behind this place it continually grows wider to the posterior 



Text-fio-ure 2. 



Dictyonosteus arcticus Stensio. 

 Transversal section through the neurocranium a short distance anterior to the 

 opticus exit. Cartilage dotted. Somewhat more than a third of the natural 

 size. 

 -Fr, frontal ; Psp/i, parasphenoid ; Sp/;, sphenoid ; ey, cranial cavitj'. 



end, Avhich is rather broad owing to the large basipterygoid 

 process (hp) on eitlier side. The posterior surface is rather 

 concave (PL I, fig. 3), probably in the same way as Watson 

 (1921, p. 335) has found it to be in Megalichthys and in the 

 Carboniferous and Permian Coelacanthids. In my description 

 in 1918, when the specimen had not been cleaned from the 

 matrix*, I came to the incorrect conclusion that this concavity 

 was the fossa hypophyseos and the upper boundary of it the 

 dorsum sellse (cf. Stensio, 1918 5, p. 120; pi. iv. fig. 3; pi. v. 

 fig. 1 ; pi. vi.). 



The ventral surface of the pars basalis is straight and plane, 

 and has been situated almost horizontally ; the dorsal one, on the 

 other hand, slopes rather much forwards, and is concave from one 

 lateral side to the other — i. e., it is somewhat grooved in a 

 longitudinal direction. At fibout the middle of its length the 



* The cleaning of the specimen from the matrix has heen verj' difficult, and has 

 taken a long time, as the matrix was a very hard ironstone. 



