Having left th e pterotic, th e canal traverses a groove in th e lateral edge of th e lateral extrascapular,
b u t no organ could be found related to th a t bone. The canal then traverses in succession the suprascapular
and supraclavicular, in each of which bones there is a single organ.
The main infraorbital canal of Peristedion thus differs from th a t of Trigla in th a t th e pterotic
lodges b u t one organ innervated by the oticus lateralis, instead of two, without intervening primary
tu b e ; and in th a t there is no apparent organ related to the lateral extrascapular.
The supratemporal canal lodges, as in all the other fishes of th e group, two organs, one lying
in the lateral extrascapular and th e other in the parieto-extrascapular.
The supraorbital canal agrees strictly with the canal in Scorpaena and Trigla, b u t between
the nasal and frontal bones th e canal is enclosed, for a relatively long distance, in the dermal portion
of the mesethmoid bone; this section of canal lodging no sense organ. The nasal lodges a single sense
organ, and th e frontal five organs, th e fourth and fifth organs pf th e line lying close together without
intervening primary tube, as in Scorpaena and Trigla. The fourth primary tubes of opposite sides
unite in the middle line to form a frontal commissure.
The preoperculo-mandibular canal contains ten organs, instead of eleven as in Trigla, three
of these organs lying in the dentary, one in th e articular and six in the preopercular. After leaving
the dorsal end of the preopercular th e canal traverses a short suprapreopercular bone, without related
organ, to reach and anastomose with the m ain infraorbital canal between th e two organs in th e pterotic.
The primary tubes, in Peristedion, do not branch repeatedly after entering th e dermal tissues,
as they do in Scorpaena, Cottus and Trigla. Certain of th e tubes, however, apparently undergo one
or two subdivisions, and th e mouths of all of them become enlarged and most of them are closed
by drum-head-like membranes perforated by several small openings.
IV. Dacty 1 opterus vol i tans.
1. S K U L L .
Of this fish I have had five specimens, ranging from 13 cm to 41 cm in length, and a number
of specimens ranging from 5 cm to 10 cm in' length. A single one of these specimens, 35 cm in
length, was th e only one I had during th e earlier periods of the investigation, the other-large specimens
being later obtained, one a t a time. The study of th e cranial bones presenting peculiar difficulties,
all of these large specimens were successively sacrificed to it, th e soft pa rts thus not being examined
in any of th e adults. When the small specimens were later obtained, several of them were sectioned,
but, because of th e character of the bones and other tissues, only one comparatively good series of
sections was obtained. This series alone served for the study of th e nerves.
Gill (’90, p. *245) says, of the skull of Dactylopterus, th a t „its upper surface is derived partly
from a dermal ossification which is incongruous with th e tru e bones“ ., J u s t what this statement
means is not clear, for th e bones in this fish differ in no way, excepting in degree, from those in Trigla
and Peristedion. As in these two la tte r fishes, all of th e primary bones th a t come to the level of,
and form part of the dorsal surface of the skull of Dactylopterus have an external surface exactly
similar to th a t of the adjoining and purely dermal bones; and the sections of young Dactylopterus
show, b e tte r even th a n those of the other fishes of the group, th a t this external portion of these
primary bones is formed b y osseous accretions th a t are apparently developed in exactly the same
way as th e corresponding portions of the purely dermal bones. A 13 cm specimen was treated with
chlorine, in -an early a ttem p t to trace the sutures between th e cranial bones, and m this preparation
a superficial layer of bone could be stripped off from both the purely dermal and the primary bones,
leaving, in the former case, a th in remaining plate of bone which may perhaps represent a separate,
membrane component underlying a more important dermal or latero-sensory component. This
membrane component would then be th e p a rt th a t persists in those higher animals in which the
latero-sensory component has disappeared.
The skull of D actylopterus is said by Cuvier & Valenciennes (’29, vol. 4, p. 131) to be depressed
and widened in such a manner th a t it represents a subrectangular disk, th e anterior edge of which
is curved in an obtuse angle, and its posterior angles prolonged into long points. A very large median
ethmoid and two prefrontals are said to form an anterior row of the bones th a t form the pavementlike
dorsal surface of th e skull. A second row is said to be formed by th e large frontals, behind each
of which bones there is a small postfrontal. A th ird row is formed by the median interparietal, the
two parietals, and th e two mastoids; and a fourth row by th e two external occipitals and two
suprascapulars. Between the third, and fourth rows, on either side, two oval bones are said to be
intercalated, these two bones together, on each side, representing th e ,,rocher“.
The prefrontals of this terminology are the ectethmoids of th e nomenclature employed by
me, th e interparietal is the supraoccipital, and the mastoid is th e pterotic. The term .,rocher“ , as
used by earlier authors, is said by Starks (’01) to be th e synonym of the opisthotic of later authors;
b u t it will be shown th a t th e so-called „rocher“ of Dactylopterus is th e lateral extrascapular, and
no t th e opisthotic of the fish, this la tte r'b o n e being wholly absent. I t will be further shown th a t
th e external occipital is a mesial extrascapular, and n ot a n epiotic; and, what is much more important,
it will be shown th a t the median ethmoid is not an ethmoid bone a t all, b u t is a median bone formed
by the fusion, in th e middle line, of the two nasals.
Gill (’88) calls th e median ethmoid of Cuvier and Valenciennes the prosethmoid, and says
th a t i t and th e anteal (vomer) are „entirely disconnected, leaving a capacious rostral chamber opening
backwards mesially into the interorbital region“ . In to this rostral chamber the well developed ascending
pedicles of the intermaxillines (premaxillaries) are said to glide. And as Gill,tin his descriptions
of the Loricati, says th a t th e ascending pedicles of th e intermaxillines glide „over the front of the
prosethmoid“ , th e term prosethmoid, as used in his descriptions of Dactylopterus, is certainly intended
by him to designate a bone th a t he considered as th e stric t homologue of th e median, primary
mesethmoid of current descriptions. No mention is made of a rostral either by Gill or by Cuvier
and Valenciennes.
The floor of the rostral chamber, in my medium-sized specimens, inclines downward and
forward a t an angle of about 45°. In the 41 cm and 35 cm specimens this floor is somewhat less steep,
while in th e small specimens it is much steeper. In the middle line of the anterior half of the floor,
there is a narrow strip of cartilage formed by a median rod-like projection from the anterior end
of the antorbital cartilage. This rod-like process is the stric t homologue of the prenasal process,
or beak, of my descriptions of Amia and Scomber, b u t it is here a relatively long and narrow rod,
curving downward and forward. I t lies in a median groove on the dorsal surface of the vomer, and
extends to the anterior edge of th a t bone, agreeing in this with the arrangement found in Scomber.