245 
the cranial cavity. The formation, here, of two floors to the cranial 
cavity instead of one seems, in fact, to be wholly due to a relative 
acceleration in the formation of the basioccipital, the earlier layers 
of that bone developing and enclosing the ventral cartilaginous ver- 
tebral processes before those processes have fused with the dorsal 
processes. Similar dorsal and ventral cartilaginous bands are shown 
by SEWERTZOFF (1902) in an embryo of Ceratodus, and these epichordal 
and hypochordal bands of that author’s descriptions are said by him 
to form, respectively, a secondary and primary floor to the cranial 
cavity; but the space between these two floors must open externally 
only, as it does in Lepidosteus, for in the adult Ceratodus there is no 
slightest indication of a related recess in the cranial cavity (GUNTHER, 
rows Pl. 35, Pigs 2). 
In Scomber (Aruıs, 1903) and Scorpaena (Auuıs, 1909), because 
of a relative diminution in the length of that portion of the neuro- 
cranium that lies between the foramina for the nervi opticus and 
trigeminus, the pituitary fossa has become relatively less long, antero- 
posteriorly, than it is in either Chlamydoselachus or Lepidosteus, 
and the membrane that roofs the anterior half of the fossa is more 
definitely differentiated. A median pit in the membrane projects 
ventrally into the fossa and lodges the hypophysis and the ventral 
end of the infundibulum, and in Scomber the saccus vasculosis also. 
The lobi inferiores do not project into the fossa remaining in the 
cranial cavity proper and resting upon ossified portions of the roof 
of the fossa. The orbital opening of the fossa, which is simply the 
enlarged pituitary foramen of Chlamydoselachus and of Lepidosteus, 
has acquired a sort of atrial space, this space being bounded by a 
floor and a lateral wall which quite certainly correspond to the floor 
and lateral wall of the trigemino-pituitary fossa of selachians. For 
these two walls of this atrial space would evidently be produced if 
the floor and lateral wall of the trigemino-pituitary fossa of Chlamydo- 
selachus were to remain in situ, held in place by their attachments 
to the wall of the epiotic capsule, when the thick interorbital wall 
of the fish was compressed to form the thin interorbital septa of 
Scomber and Scorpaena. The floor of this atrial space would then 
represent the basal portion of the basipterygoid process of Lepidosteus, 
and probably the entire basipterygoid process of higher vertebrates. 
And this method of origin of the basipterygoid process, both of 
Lepidosteus and of higher vertebrates, would seem to be a much more 
