brain case. The latter surface inclines slightly forward, but lies, as just above stated, in a nearly
transverse position. I t is formed, as usual, by the prootic, alisphenoid and sphenotic, but the latter
bone is reduced and forms only the small dorso-lateral corner of the wall. The wide roof of the orbit
is formed mainly by the frontal but partly also by the ectethmoid and sphenotic. Its floor is formed
in part by the expanded base of the pedicle of the ectethmoid, in part by the wide parasphenoid,
and in part also by the large orbital shelf of the second bone of the infraorbital series.
The interorbital wall is relatively thick. I t is single in its ventral but double in its dorsal
portion, this latter portion enclosing an anterior prolongation of the cranial cavity, which extends
the full length of the interorbital region and lodges the olfactory nerves. The anterior half, approximately,
of the wall is cartilaginous, and here the olfactory prolongation of the cranial cavity is roofed
by a wide flat band of cartilage, already referred to when describing the rostral chamber. The
posterior half of the wall is formed, in its ventral portion, by a median interorbital process of the
parasphenoid, and in its dorsal portion by the alisphenoid of either side, the expanded dorsal edges
of the latter bones not quite touching in the median line and so leaving a narrow longitudinal opening
in the roof of this part of the olfactory extension of the cranial cavity. A ventral flange to the frontal,
found more or less developed in all the other fishes of the group is here wholly wanting.
The cartilaginous portion of the interorbital wall is perforated, close to its antero-dorsal corner,
by a large oval opening which leads from orbit to orbit and is closed, in the recent state, by membrane.
This membrane is single in its ventral but double in its dorsal portion, the latter portion enclosing
the anterior end of the olfactory prolongation of the cranial cavity, which prolongation extends
to the hind surface of the short pillar of cartilage th a t forms the hind wall of the rostral chamber.
The membrane is pierced, on either side, by the olfactory nerve, th a t nerve then traversing the
extreme anterior end of the orbit to enter and traverse the opening th a t leads from the orbit into
the nasal pit, th a t pit being, as already stated, confluent with the rostral chamber. From the single,
ventral portion of the membrane, ventral to the olfactory nerves, theobliqui muscles have their origins.
The median interorbital process of the parasphenoid, above referred to, is a tall broad Y-
shaped process, the spreading arms of which may be said to present three regions. In the anterior
region each arm is formed by a thin plate of bone which overlaps externally the anterior edge of
the corresponding alisphenoid, and, anterior to th a t bone, lies against the external surface of a part
of the cartilage th a t encloses the interorbital extension of the cranial cavity. In the middle region,
the dorsal edge of either arm is thickened and suturates with the ventral edge of the corresponding
alisphenoid. In this region the dorsal surface of the process seems to form the floor of the interorbital
extension of the cranial cavity, but there may here have been delicate lining plates of cartilage that
were lost in dissection. The basal portions of the arms of the Y are connected, a t the hind edge of
this middle region, by a delicate transverse web of bone. In the posterior region, the arms of the Y
are short and spread considerably, thus forming, on the dorsal edge of this part of the process, a
basin-like depression in which the optic chiasma rests. The alisphenoid of either side arches above
this optic depression of the process of the parasphenoid and, anterior to it, suturates, as already
stated, with the corresponding dorsal edge of the olfactory portion of the process. The posterior
corner of each arm of the Y is prolonged and terminates in a point directed toward, or even reaching
the dorsal edge of the orbital surface of the prootic. A large fenestration of the interorbital wall is
thus here formed which may be wholly enclosed by the bounding bones, those bones being the ali-
sphenoids and prootics of either side and the median interorbital process of the parasphenoid. The
fenestration, in the prepared skull, leads from orbit to orbit, traversing the interorbital portion of
the cranial cavity. In the recent state, the fenestra of either side is closed by a membrane which
is pierced by the optic nerve as it passes from the cranial cavity to the orbit. The two fenestrae
are accordingly the optic fenestrae.
The interorbital process of th e parasphenoid of Dactylopterus would seem to be the homologue
of th e median process of th e same bone of Gymnarchus, as shown in ErdTs (’47) figures, which
process is considered b y th a t author as the lower portion of th e ala magna, and by Ridewood (’04b,
p. 198) as th e basisphenoid. I t seems also to be th e equivalent of th e basisphenoid of Ameiurus
(’Me Murrich, ’84), fused, perhaps, with th e orbitosphenoid of th a t fish. I t is apparently the homologue
of the median process of the parasphenoid of Peristedion, b u t enormously developed. That
it contains an originally independent basisphenoid element, as its general relations would certainly
indicate, seems improbable, for even in 5 cm specimens there is no slightest indication of two independent
ossifications. In these la tte r specimens, the process is apparently wholly of membrane
bone, b u t i t is in p a rt formed by two th in laminae of bone which enclose between them a p a rt of
the cartilage of th e interorbital septum, much as th e pedicle of th e basisphenoid does in young
specimens of Scorpaena.
Ventral to th e optic fenestra, and, in most of my specimens, pa rtly confluent with it, there
is a second large perforation of th e posterior portion of th e interorbital wall, this perforation also
leading from o rbit to orbit b u t n o t traversing any portion of the cranial cavity. This perforation
is bounded anteriorly by the hind edge of th e interorbital process of the parasphenoid. Ventro-posteriorly
it is bounded by a ta ll and th in ridge of bone which extends transversely across th e dorsal surface
of th e parasphenoid between the small and pointed ascending processes of th a t bone. This
ridge of bone projects dorso-posteriorly, suturates on either side with th e anterior edge of the ventral
portion of the prootic and slightly also with the ventral edge of th e orbital portion of th a t bone, b u t
between th e two prootics presents a free dorsal edge. In th e recent state, a membrane extends from
this free portion of th e edge of the ridge upward and forward to th e concave hind edge of th e spreading
arms of th e orbital portion of the interorbital process of th e parasphenoid, th e lateral edges of the
membrane, postero-ventral to those arms, being attached, on either side, to the mesial edge of the
nearly vertical orbital portion of th e corresponding prootic. Against th a t p a rt of this membrane
th a t lies between th e orbital portions of th e prootics, or immediately postero-ventral to it, lies the
pitu ita ry body, th e entire opening closed b y th e membrane, or a t least th a t p a rt of it th a t lies between
the prootics, accordingly being th e pitu ita ry opening of th e brain case. The whole opening may be
referred to as th a t opening.
The ventral one of th e two usually confluent perforations of the interorbital wall of Dactylopterus
th u s lies between a membrane th a t fills the p itu ita ry opening of th e brain case and a process
of th e parasphenoid the dorsal end of which fulfils the function of a basisphenoid, if it b e n o t in p a rt
th a t bone. On the antero-dorsal portion of th e osseous boundary of this ventral perforation of the
interorbital wall, and pa rtly also on th e ventral surface of the membrane th a t closes th e pituitary
opening, a median vertical m embrane has its attachment. Ventro-posteriorly this membrane becomes
less strong, and separates into two parts which spread to either side and are doubtless attached to
the parasphenoid, though this could not be satisfactorily determined in my material. The membrane
is thus a median vertical one which closes, more or less completely, th e ventral perforation. On
the dorsal portion of this membrane, in 5 cm specimens, and directly opposite its fellow of the opposite