A short leg of the alisphenoid forms, in the adult, the anterior boundary of the oculomotorius
foramen, the ventral end of the leg being continuous with the small flat bar of cartilage th a t lies
between the ventral portion of the optic fenestra and the foramen for the pituitary vein; and as this
bar of cartilage rises from the lateral edge of the basisphenoid region of the trabecular cartilage, the
associated leg of the alisphenoid must be the basisphenoid leg of th a t bone. Of the parasphenoid
leg of the alisphenoid I can find no trace, the alisphenoid of Lepidosteus thus being strictly of the
usual teleostean type.
The recti muscles of the eye all have their origins from the skull immediately ventral and
posterior to the optic fenestra, and hence in the immediate neighbourhood of the foramen for the
pituitary vein. The rectus externus, in the 80 mm specimen, separates into two portions as it approaches
the skull, and the tendon of one of these portions traverses the foramen with the vein, and,
following it, has its origin beneath it; but whether its point of origin is on the cartilage of the basis
cranii, on the parasphenoid, or in fibrous tissue of the region, I could not determine.
Although wholly unrelated to th e present subject, it may here be s tated th a t the rectus in-
ternus arises, in Lepidosteus, by a long and slender tendon which is closely applied to th e antero-
mesial surface of th e rectus inferior, these two muscles thus arising as a single muscle and being,
practically, not y e t fully differentiated parts of a single muscle; this being in accord with m y supposition
(97a) th a t these two muscles of teleosts and bony ganoids arise by th e splitting of th e single
rectus inferior of elasmobranchs. These two muscles, and th e rectus superior and obliquus inferior,
are innervated b y the oculomotorius in the manner and order th a t they are in Amia, this confirming
my interpretation (’97a, p. 520) of Van Wijhe’s description of this fish. The abducens leaves the
cranial cavity through the trigemino-facialis foramen, piercing the membrane th a t closes th a t foramen
antero-ventral to and independent of th e trigemino-facialis roots. I t then runs antero-laterally, v en tra
l to the jugular vein, between it and th e external carotid, lying internal and antero-internal to
th e ventral edge of the trigemino-facialis ganglion, and reaches th e rectus externus. I t does not
pass over th e truncus trigeminus, as s tated by Schneider (’81), but, as the rectus externus passes
over th a t truncus, the abducens naturally also would, if it were sufficiently prolonged. The radix
profundi, after its exit from the skull, enters an extra-cranial profundus ganglion which lies close
against th e side wall of th e skull immediately dorso external to and in contact with th e oculomo-
torius, and n o t below th a t nerve, as shown by Schneider. Two ventra l or ventro-anterior prolongations
of the ganglion terminate in small ganglionic swellings, which lie in contact with the inferior
branch of the oculomotorius and are th e ciliary ganglia. From the anterior end of th e profundus
ganglion a stout portio ophthalmici profundi arises. The trigemino-facialis ganglion lies, as in Amia,
wholly outside th e cavum cranii, a few scattered cells only being found in the roots of th e nerves.
The cross-canal of Sagemehl’s descriptions can now be considered. The enclosing walls of
this canal form, in the fresh skull of the ad u lt fish, a prominent transverse bolster on the floor of
th e cranial cavity, and this bolster must quite unquestionably have arisen by th e depression of the
anterior edge of th e prootic bridge of a fish like Amia (Allis, ’97a, Fig. 11) or Polypterus (Pollard,
’92, Fig. 12) until it met and then fused with th e floor of the cranial cavity on either side of the
saccus vasculosus. The cross-canal of Lepidosteus is thus a strictly intramural space th a t
lies beneath a prootic bridge of th e primary type found in Amia, and th a t is related either primarily
or secondarily to the saccus vasculosus. This space represents th e posterior half only of the myo-
dome of Amia, and it might have been developed either from th e conditions found in' th a t fish, which
has a myodome, or from those found in Polypterus which has no myodome. In Lepidosteus the
mesial processes of the proôtics, which unite to form the prootic bridge, are said b y Sagemehl to be
connected, across the median line, by membrane, instead of by cartilage. To me this connecting
tissue looks much more like cartilage or fibro-cartilage th a n like membrane, b u t this is unimportant,
for even in Amia membrane is here first formed and later chondrifies. Sagemehl further says th a t
this membrane entirèly closes the canal toward the cranial cavity. This is an error, for I find the
anterior wall of th e canal always perforated b y a median oval opening through which the apparently
short saccus vasculosus projects. The cross-canal is filled, as Sagemehl states, with a mass of fa tty
tissue and this tissue is richly supplied with blood.
Immediately anterior to the cross-canal there is, in th e floor of th e cranial cavity, the p itlike
and slightly oval pitu ita ry fossa. The lateral and anterior walls of this fossa are of cartilage.
The floor of th e fossa is perforated by a nearly circular opening, the opening being closed ventrally
by the parasphenoid. Posterior to this circular opening, and extending to or slightly beyond the
anterior wall of th e cross-canal, the floor of the p it is, in my medium-sized adults, formed of tough
membrane only. Posterior to this membrane th e floor of the cross-canal is of cartilage, and inclines
gradually and slightly upward to the base of the ridge of cartilage th a t forms the posterior boundary
of th e cross-canal. The circular opening in the cranial floor is accordingly the hypophysial fenestra,
th e greatly reduced pitu ita ry space, or pitu ita ry fenestra of Parker’s descriptions of embryos of
Lepidosteus, and this fenestra of th e ad u lt Lepidosteus is the strict, b u t reduced homologue of the
hypophysial fenestra of Amia. The fenestra of Lepidosteus is said by Parker to lie between the
„roots“ of th e trabeculae; but, as already stated, the laterally bounding cartilages seem to me to
be th e anterior ends of th e parachordal cartilages. Parker’s opinion was doubtless based; 1st., on
th e position of the hypophysis, in early embryos, between th e hind ends of th e trabeculae; and, 2nd.
on the position of th e posterior clinoid bridge of his descriptions, which bridge is said by him (1. c.
p. 481) to run „straight across, joining th e roots of th e trabeculae together“. But, this posterior
clinoid bridge simply represents an early stage in the development of the enclosing walls of Sagemehl’s
cross-canal, and hence joins regions of parachordal and not of trabecular origin. Furthermore,
th e position of the hypophysis in early embryos need not necessarily be th e same as th a t in the adult.
For, the intertrabecular and interparachordal fenestrae of early embryos being continuous, as shown
by Parker, th e well-known unequal longitudinal growth of the cartilage of the basis cranii and the
overlying brain might easily pull the pitu ita ry body backward, out of the hind end of the inte rtrabecular
space into the interparachordal region. And this, in my opinion, has certainly taken place
in Lepidosteus.
Sagemehl says th a t the lateral ends of the cross-canal closely approach the postero-ventral
corner of the orbit. This is an error, for the cross-canal is everywhere widely separated from the orbit.
He further says th a t the cross-canal exactly resembles the canalis transversus of selachians, excepting
th a t it is closed toward th e orbit. This also is not wholly correct, for as the canalis transversus of
selachians is traversed (Allis, ’01) by a venous sinus and not simply by a lymph sinus, as Sagemehl
supposed, it must be represented, in p a rt a t least, in Lepidosteus, by the canal th a t transmits thé"
p itu ita ry vein. There are here, in fact, two spaces th a t would seem to be of separate and independent
origin; one a median space th a t is either primarily or secondarily related to th e saccus vasculosus, and
the other a canal th a t leads from the orbit, on either side, and is traversed by th e pituitary vein. The
saccular space is roofed by ¿he mesial processes of the proôtics, and from it, or immediately in front