The intermuscular system of acanthomorph fishes : a commentary

24 Pages·2001·3.5 MB·English

by  Johnson, G. David

AMERICAN MUSEUM
Novtitates
PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY
CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024
Number 3312, 24 pp., 6 figures January 30, 2001
The Intermuscular System of Acanthomorph
Fishes: a Commentary
G. DAVID JOHNSON! AND COLIN PATTERSON?
We got some ‘splainin’ to do.
Desi Arnaz, as Ricky Ricardo in “I Love Lucy”
You can’t have your cake and eat it too.
Ancient Proverb
ABSTRACT
This paper is a response to Gemballa and Britz (1998), who presented a new interpretation
of the intermuscular bones of acanthomorphs, homologizing them with the epicentrals of lower
teleosts. We argue that their identification of epineural ligaments above the intermuscular bones
in many acanthomorphs is mistaken; the structures in question are fanlike arrays of collagen
fibers, not true intermuscular ligaments. We show also that undisputed epineural intermusculars
penetrate or enter the horizontal septum in lower acanthomorphs (Velifer, Polymixia, beryci-
forms), and reiterate arguments for regarding the single series of intermusculars in most acan-
thomorphs as epineurals, secondarily displaced into the horizontal septum.
‘Curator, Division of Fishes, National Museum of Natural History, Smithsonian Institution, Washington DC
20560.
? Honorary Research Fellow, FRS, deceased 1998, Department of Palaeontology, The Natural History Museum,
London.
Copyright © American Museum of Natural History 2001 ISSN 0003-0082 / Price $4.00
2 AMERICAN MUSEUM NOVITATES NO. 3312
INTRODUCTION eton and thus act as tendons’’; (3) when in-
termuscular bones develop within the struc-
Gemballa and Britz (1998; hereafter re-
tures, thereby justifying the term ligament for
ferred to as “G & B’’) have published an
the proximal part joining bone to axial skel-
account of the intermuscular bones and lig-
eton, the bone appears relatively late in on-
aments in acanthomorphs. Their interpreta-
togeny, “‘and we see no necessity to apply a
tion of those structures differs in several
different term to the same structure at a later
ways from ours (Johnson and Patterson,
point in ontogeny.”
1993; Patterson and Johnson, 1995; hereafter
We regret and will oppose their decision
referred to respectively as “J & P”’ and “P
to change the names, in part because the in-
& J’’). In brief, G & B’s main conclusion is
termuscular system is already difficult, and
that the single series of intermuscular bones
to change the entire terminology only three
found in the vast majority of acanthomorphs
years after the system was first mapped (in
are not epineurals displaced ventrally into the
P & J) will not help communication. We also
horizontal septum, as we proposed, but are
question the propriety of using tendon for all
epicentrals. [Incidentally, this was also the
the structures that we called ligaments. The
opinion of Richard Owen, who first named
technical distinction is ambiguous in a num-
the three intermuscular series (1866), and be-
ber of cases. The term tendon was used for
lieved that the bones in Perca and Gadus,
structures in the horizontal septum by Kish-
the acanthomorphs in his sample, were epi-
inouye (1923) and Kafuku (1950), who
centrals. As we wrote (P & J, p. 2), ““we both
named the anterior and posterior oblique ten-
initially thought Owen right.”?] G & B found
dons (AOTs, POTs); Kafuku’s AOTs are our
that our conclusion resulted from inadequate
epicentral ligaments (in lower teleosts) and
technique, which led us to miss ligamentous
epineural ligaments (in percomorphs). Ka-
elements of the intermuscular system in
fuku named these structures tendons because
many of the acanthomorphs that we exam-
they connect the lateralis superficialis (red
ined. The purpose of the present paper is to
muscle) with the axial skeleton. Westneat et
explain some of the differences between their
al. (1993) gave a detailed account of the ar-
observations and our own and the _ back-
rangement of AOTs and POTs in the hori-
ground to our conclusions; then to review the
zontal septum of scombrids and their role in
different interpretations of structure in some
the transmission of forces to the axial skel-
acanthomorph taxa; and finally to compare
eton. Among other things, they found that
the two arguments for homology. In the In-
AOTs and POTs are joined distally into a
troduction to our paper (P & J) we wrote that
loop by a segment of the POT that they
our observations of ligaments “include an
called ITL (intertendon length)—‘“‘This loop
unknown quantity of subjectivity, and we
is so robust that the fish can be lifted by it.”
will be glad to see all or any of them checked
Each POT slides relative to succeeding AOTs
by others.’’ We did not anticipate that we
as it passes posteriorly, but a significant por-
should so soon be required to check some of
tion of its fibers converge with and join those
them ourselves. To anticipate our results, af-
of the distal end of the anteriormost AOT of
ter a lot of work and some initial doubts, we
the loop (Westneat et al., 1993: 195). Here
are convinced that the intermuscular bones of
then, AOT and POT together form what is
acanthomorphs are epineurals.
technically a ligament, since it joins two
bones (vertebrae). Also, although POTs gen-
TERMINOLOGY
erally insert in the lateralis superficialis, the
Following advice from M. W. Westneat, G anteriormost POTs may insert on intermus-
& B elected to use the name “‘tendon’’ for cular bones in percomorphs (P & J), here
the connective tissue bundles in myosepta again meeting the technical criterion of lig-
that we (P & J) named “ligaments.”’ Their ament.
reasons for the change of name were: (1) the The other series of structures that we (P &
structures connect muscle to bone (axial J) named ligaments are the epineurals and
Skeleton); (2) the structures ‘“‘transfer mus- epipleurals. Members of these series have
cular forces from myomeres to the axial skel- been best illustrated by Gemballa (1995).
2001 JOHNSON AND PATTERSON: ACANTHOMORPH INTERMUSCULARS 3
They attach to the axial skeleton proximally (which he called intramuscular) bones (spe-
and lie in the myoseptum, where they fan out cifically epineurals), and also used the term
distally into an array of small fiber bundles ligament for unossified intermuscular ele-
that have no specific insertion in the body ments. Perhaps reinforcement and stabiliza-
musculature. Some epaxial muscle fibers at- tion do not exclude a transmission of force
tach perpendicularly along the length of the function as well, but if the latter is the cri-
intermusculars (be they ossified or unossi- terion by which we are to distinguish tendon
fied), as they do over the entire surface of from ligament, then structures such as the
the myoseptum, but the muscle fibers do not maxillo-mandibular ligament, semicircular
converge or concentrate at the points of in- ligament, and others must also be renamed.
sertion as is typically the case for tendons. G G & B’s usage also entails some ambigu-
& B stated that such fibers “‘transfer mus- ity. For example, in describing Velifer they
cular forces from myomeres to the axial skel- called the ligaments attaching the epineural
eton and thus act as tendons”’ but this is to bones of V7—10 ‘‘a short proximal tendinous
oversimplify, and of course the same then sheath.”” We will therefore continue to use
must be true of the entire myoseptum, as dis- our earlier terminology, and call the connec-
cussed by Van Leeuwen (1999). Further- tive tissue bundles ligaments, not tendons.
more, we are unable to reconcile G & B’s
conclusion (nor did they address it) about
Abbreviations
transmission of forces with the following
statement from Gemballa (1995: iii), in BL Baudelot’s ligament
cf collagen fiber
which the term ligament was used for the
ecb epicentral bone (G & B only)
unossified elements of all three series:
ecc epicentral cartilage
ecl epicentral ligament
Epineural and epipleural ligaments are unlikely to
play a role in transmission of muscular forces because enb- epineural bone
its collagen fibres and white muscle fibers include an ent epineural tendon (G & B only)
obtuse angle. They resist the radial expansion of con- epb- epipleural bone
tracting muscle fibres of the VAC’s and, DAC’s and Ib lateral band (G & B only)
thus may stiffen the trunk during undulatory loco- ns neural spine
motion (bracing and bandaging “‘Querverspannungs- Vv vertebra
funktion’’). Furthermore during sharp bendings these
ligaments keep the relative positions of axial skeleton Institutional abbreviations are those of
and integument to each other against the expanding
Leviton et al. (1985).
musculature (holding device or guy, ‘“‘Haltefunk-
tion’’).
DIFFERENT METHODS; DIFFERENT
In a more recent paper, Gemballa (1998:
CRITERIA; DIFFERENT RESULTS?
30) hypothesized that ‘““Muscle arches con-
sisting of white muscle fibers are reinforced G & B used the following technique de-
by epineural tendons and lateral bands that veloped by Gemballa during work for his
may (emphasis ours) transmit forces to the doctoral thesis (1995): cleared-and-stained (c
backbone.”’ It would appear that the precise & s) specimens are transferred from glycerin
function of unossified epineural and epi- to ethanol; then individual myosepta are cut
pleural elements remains unresolved. Epi- free, mounted on slides, and studied with po-
neural and epipleural intermuscular bones, larized light. With this method, G & B are
when present, are ossifications in the epineu- able to visualize the collagen fibers in a myo-
ral and epipleural ligaments. G & B’s state- septum far more accurately than we could,
ment, quoted above, means that the function with our relatively primitive technique of ob-
of these intermuscular bones is to transmit serving c & Ss specimens under the stereo-
the pull of muscles to the axial skeleton, but microscope. Although we mentioned the val-
Wainwright (1983: 70) stated that “Inter- ue of transferring glycerin specimens to al-
muscular bones, when present, reinforce sep- cohol (P & J: 4), we made little use of the
ta,” and Van Leeuwen (1999), in his me- technique. This was in part because we felt
chanical analysis of myomere shape, referred that the procedure, like too much reliance on
to the stabilizing effects of intermuscular incident light, could cause us to record as
-: AMERICAN MUSEUM NOVITATES NO. 3312
ligaments aggregations of collagen that did HISTORICAL ACCIDENT
not deserve the name. We were anxious not
As chronicled in the Introduction, our
to Overinterpret the specimens. In our exten-
study of the intermuscular system began with
sive survey of teleostean intermusculars (we
an effort by one of us (CP) to understand the
recorded the system in about 125 genera) we
system in a large cleared-and-stained Poly-
had to develop a criterion for what should be
mixia where the epicentral series cannot be
called an intermuscular ligament and what
missed because the ligaments of V3—17 con-
should not. This is always necessary toward
tain conspicuous cartilage rods. We agreed to
the caudal end of each series, where the lig-
collaborate after GDJ had seen a draft man-
aments become successively shorter and
uscript, found that intermuscular ligaments
more subdivided distally, and is often nec-
are far more widely distributed than CP had
essary also at the cranial end of a series.
thought, and carried to London a number of
Working in untrodden ground, we identified
c & s specimens. Notable among these were
ligaments by trial and error, changing places
holocentrid and anomalopid beryciforms that
at the microscope or passing specimens from
seemed to show a continuous series of inter-
hand to hand between two microscopes. We
muscular bones and ligaments of which the
arrived at the following criterion of discrete-
first few members are in the horizontal sep-
ness: we recorded as ligaments only those
tum (the usual position of epicentral bones
structures that, like ossified intermusculars,
and/or ligaments), the last few on the neural
had a single origin on the axial skeleton, and
spines (the usual position of posterior epi-
were sufficiently discrete to be picked up dis-
neurals), and the intervening members as-
tally by the forceps (recognizing that the lat-
cend successively further from the horizontal
ter criterion can have an element of subjec- septum. Beneath some of those ascending
tivity and is difficult to apply in the very
bones, and beneath all of the posterior ones,
smallest specimens). Where a series of bones
is a series of epicentral ligaments, but that
is continued rostrally or caudally by lga-
series seemed to us incomplete because none
ments our “‘discrete’’ structure characterizes
occurred on the first few vertebrae. Polymix-
those ligaments that lie immediately adjacent
ia, on the other hand, has only the first in-
to the bones with which they are unquestion- termuscular bone in the horizontal septum,
ably serially homologous. With this criterion, those of succeeding vertebrae successively
we would not have recorded as ligaments further above that septum, and an epicentral
many of the structures so labeled by G & B. series of ligaments that extends forward to
This is true notably of their first epineural
V2. These fishes, interpreted as a morpho-
ligament (enl 1), which they label epineural cline continued with further modification in
tendon (ent) in Polymixia (their fig. 2B, our percomorphs, were the key to our interpre-
fig. 1A), and of the epineural ligament (ten- tation of acanthomorph intermuscular bones
don) in the following illustrations: their fig.
as descended epineurals. Later in our inves-
4B-D (ent 6, 3, 1 in Hoplostethus), fig. 6A tigation of percomorphs, we found (incon-
(ent 15 in Aphredoderus), fig. 6B (in the veniently) that some of them had a series of
midbody of Molva), fig. 6C (ent 15 in Be- ligaments above the intermuscular bones, in
dotia), fig. 6D (in the midbody of Scomber- the position of the epineural series of lower
esox), fig. 6E, F (ent 8, 2 in Lates), fig. 7A, teleosts. We had to interpret these as neo-
B (ent 9, 3 in Centropomus), fig. 7C, E—-F morphs, and called them neoneurals. We re-
(ent 11, 2, 1 in Morone) fig. 8A—C (ent 15, corded them in representatives of almost 20
8, 2 in Serranus, fig. 8D, E (ent 8, 3 in Pseu- families (Ammodytidae, Bothidae, Caesion-
danthias), and fig. 1OE (ent 19 in Channa). idae, Carangidae, Echeneidae, Gerreidae,
We question also the epicentral ligaments Haemulidae, Kuhliidae, Labridae, Lethrini-
(ecl) indicated in G & B’s fig. 6E (ecl 8 in dae, Lutjanidae, Mullidae, Polynemidae, Po-
Lates) and fig. 7A (ecl 9 in Centropomus). macentridae, Samaridae, Scaridae, Sciaeni-
We might argue about a few more of the il- dae, Sparidae, Teraponidae; they are ossified
lustrated structures, but we comment in detail in the bothid and samarid pleuronectiforms
below on those in the above list. we studied). Naturally, G & B took our
2001 JOHNSON AND PATTERSON: ACANTHOMORPH INTERMUSCULARS E)
**neoneurals”’ to be the true epineural series, REVIEW OF STRUCTURE
and used them to counter our argument of
Here we review the intermusculars in
homology, based on interpretation of lower
some of the acanthomorphs described by G
acanthomorphs.
& B (we treat the taxa in the sequence that
However, as we extended our investigation
they used), and in a few other taxa recorded
into lower teleosts, we found a situation in
by P & J but not by G & B. We also com-
aulopiforms that seemed to bolster, if not
ment on the POTs in these fishes, because
confirm, our argument that acanthomorph
they are relevant to our argument about ho-
epineurals have descended into the horizontal mologies of acanthomorph intermusculars.
septum, displacing the epicentrals. In aulo-
piforms the epineural series is normal, and
POLYMIXIIDAE
there is no reason to question its homology
As noted above, our investigation began
with that of other lower teleosts. In the hor-
izontal septum of almost all aulopiforms with Polymixia, and it was one of the keys
to our interpretation of percomorph inter-
there is a series of bones that extends for-
muscular bones as descended epineurals. It is
ward to V1 or V2. We first took these to be
epicentrals, which ossify in a few other also the key taxon in G & B’s reinterpreta-
tion, which is based on Patterson’s (1982)
groups (e.g., Megalops, clupeomorphs, gon-
orynchiforms). But then we found aulopi- “conjunction test’? of homology: homology
between two structures (such as the wing of
forms in which the posterior ‘“‘epicentrals”’
a bird and the arm of a mammal) is refuted
moved successively further below the hori-
if an organism (such as an angel) exhibits
zontal septum until they occupied the ventral
both. G & B argued that the intermuscular
position normal for epipleurals, the third se-
bone of V1 in Polymixia, which inserts in a
ries of intermusculars. Some aulopiforms,
socket at the base of the neural arch, is larger
notably a larva of Scopelarchoides (P & J:
than its bony successors and (unlike them)
fig. 10) and the evermanellids (fig. 3),
lies in the horizontal septum, cannot be an
showed clearly that this ventral shift from the
epineural (and so must be an epicentral) be-
horizontal septum was accompanied by suc-
cause there is an epineural ligament above it
cessively deeper distal subdivision, of liga-
in series with the epineural bones (and liga-
ments (in larvae, P & J: fig. 10), or of bone
ments) of succeeding vertebrae. As also not-
with ligament attached dorsally (notosudids),
ed above, we would not have called their epi-
or of bone ventrally and ligament (sometimes
neural | (fig. 1A; G & B, fig. 2B) a ligament:
with included cartilage, ecc in fig. 3) dorsal-
it is not a single, discrete structure, but an
ly. Here we had evidence of continuity be-
array of bundles of collagen fibers that have
tween a series of bones in the horizontal sep-
multiple origins on the neural spine and de-
tum anteriorly and well below it posteriorly, crease in size dorsally and ventrally from the
and in which identification as epipleurals
larger bundles in the center of the array. Fur-
seems incontrovertible, because the epicen-
thermore, G & B’s depiction (fig. 1B; G &
tral series can be identified as the ligaments
B, fig. 2C) of the origin of that structure in
inserting on the dorsal surface of the inter-
the usual position of the first epineural, 1.e.,
muscular bones in the zone of transition.
at the base of the neural arch near its attach-
With this evidence that it was possible for ment to the centrum, is inaccurate. Their il-
epipleurals to “‘migrate’’ dorsally into the lustration shows the collagen fibers attaching
horizontal septum, we were more comfort- along an area from the dorsal margin of the
able with our interpretation of percomorph proximal tip of the first intermuscular bone
intermusculars as epineurals that had done (their ecb) to a point above its insertion on
the opposite trick, “‘migrating”’ ventrally into the base of the neural arch below the dorsal
the horizontal septum. Holocentrids and an- margin of the centrum. As shown in their
omalopids were our acanthomorph equiva- figure 2B (our 1A) and more clearly in our
lent of aulopiforms like that in figure 3, dem- figures 2, 4, and 5, the lowest point of at-
onstrating coexistence of the two series in the tachment of these collagen fibers is well
same segment. above the insertion of the first intermuscular
6 AMERICAN MUSEUM NOVITATES NO. 3312
bone (and the dorsal margin of the centrum) table 1), and we have since confirmed our
and they extend along the neural arch dor- observations.
sally to near the base of the neural spine. We used Gemballa’s technique of transfer
These observations do not refute unequivo- of glycerin specimens to alcohol followed by
cally the identity of this structure as an epi- dissection and polarization of individual
neural, however G & B’s interpretation of the myosepta to study and photograph the first
first vertebra in Polymixia requires that the two myosepta of Polymixia (fig. 2) and Ve-
first epineural has been displaced dorsally lifer (fig. 6). Our technique differed only in
from its primitive site of insertion (which of- that we used a stereomicroscope, so that our
ten has a modified articular depression or photographs are of intact epaxial myosepta,
even a well-developed socket, as in Velifer) whereas G & B used a compound micro-
and reduced from a robust bone to a diffuse scope, and reassembled several photographs
array of collagen fibers, while the first epi- to cover the entire myoseptum. We agree that
central has ossified and moved dorsally to Gemballa’s technique provides a superb way
occupy the usual insertion point of the epi- to visualize collagen fiber in myosepta, how-
ever our figures 4 and 5 of two intact c & s
neural. More important, as shown in figures
Polymixia in ethanol show that removal and
2,4, and 5, G & B’s argument may be turned
polarization of individual myosepta are not
around, because on the neural spine of V2
essential for accurate observation of inter-
there is an exactly similar, though slightly
muscular ligaments or the arrays of collagen
less extensive, array of collagen bundles. Be-
fibers (cf) identified as such by G & B. It
neath this is the intermuscular bone, which
does, however, prove informative in an un-
originates on the neural arch, lies above an
expected way. Note that in figure 2 of the
epicentral ligament, and is inclined dorsolat-
polarized myoseptum 2 in Polymixia, bone
erally like its successors. This must therefore
(enb) and true ligament (BL and ecl) are es-
be an epineural, yet above it is a “‘ligament”’
sentially indistinguishable and look nothing
like that which G & B identify as epineural
like the arrays of collagen fiber (cf) above
1. We believe that this refutes G & B’s con-
each bone.
clusion that “‘there is strong and unambigu-
Notice also in figure 2B that the epicentral
ous evidence that the first vertebra of Poly-
ligament originates not on the vertebra but
mixia has an epineural tendon [ligament].”’
on the base of the epineural; we had not ob-
Presence of a diffuse array of collagen bun-
served this in intact c & s specimens, but
dles in the myoseptum above an undoubted
have checked it by removing myoseptum 2
epineural bone in myoseptum 2 of Polymixia
from the large Polymixia illustrated in P &
also throws into question G & B’s identifi-
J’s figures 1 and 2. This means that the prox-
cation of similar arrays as epineural liga-
imal end of epineural 2 must penetrate the
ments in other acanthomorphs (e.g., their
horizontal septum (which contains the epi-
figs. 7, 8).
central), and is evidence that epineurals may
Our argument (P & J: 33) that the first
enter that septum.
intermuscular bone of Polymixia is a de-
In Polymixia (P & J: fig. 3) we found the
scended epineural, not an epicentral, was
first POT to originate on V8 and insert on
supported in our view by the fact that in a the epicentral ligament of V5.
12 mm larva the bone is fully formed and
develops in series with the epineurals (those
LAMPRIDIFORMES
of V2-6 are ossified), before the epicentral
ligaments are recognizable. G & B “‘consider VELIFERIDAE: We no longer have access to
this argument as invalid because the identi- the specimens of Metavelifer and Velifer that
fication of epicentral tendons [ligaments] in we recorded (Patterson and Johnson, 1995:
12 mm long cleared-and-stained specimens table 7). G & B do not differ from us in their
seems impossible with their technique.’’ Yet interpretation of the intermuscular bones as
with that technique we had no trouble in ob- epineurals. The only difference between their
serving a long series of epineural and epi- observations and ours concerns the epicentral
pleural ligaments in the specimen (P & J: ligament of V1: G & B failed to find one.
2001 JOHNSON AND PATTERSON: ACANTHOMORPH INTERMUSCULARS 7
We recall our own doubts about that struc- anterior) POT as originating on V9 and in-
ture. (Our specimen had been much dissected serting on the epicentral ligament of V3. G
before the drawing in J & P, fig. 1C, was & B’s (fig. 3) illustrations of Velifer show
made.) POTs on the epicentral ligament of V11, but
We have removed myosepta 1 and 2 from none are recognizable in their photographs of
a new specimen of Velifer (fig. 6). Epicentral myosepta 9 and 4.
ligament | is slender and (on both sides of LAMPRIDIDAE: G & B found that the first
the specimen) originates on the epineural two intermuscular bones of Lampris origi-
bone rather than on the centrum. As with the nate on the centrum rather than the neural
origin of epicentral 2 in Polymixia (above), arch, where we placed them. Our observa-
this means that the proximal part of epineural tions were based on two large dried skele-
1 enters the horizontal septum. Above the tons, USNM 271011 and 273477. The first
epineural in myosepta 1 and 2 there is an neural arch is fused to the centrum in Lam-
array of collagen fibers corresponding to pris, and has an unusual configuration be-
those in myoseptum 1 and 2 in Polymixia, cause the exoccipital portions of the occipital
but inserting distally on the epineural. Our condyle extend dorsally to enclose the fora-
picture of myoseptum 1 in Velifer (fig. 6A) men magnum (Olney et al., 1993: 157). Our
is very different from that in G & B (fig. belief that the facet or groove for the inter-
3D), and we have confirmed our observa- muscular is on the neural arch is confirmed
tions on the other side of this specimen and by the 19 mm larva figured by Olney et al.
both sides of an additional specimen from the (1993: fig. 7), in which the first epineural
same lot. The portion of the septum that in originates on the neural arch. On V2, where
our specimen contains the epicentral liga- the neural arch is also fused to the centrum,
ment is filled with a tangle of fibers in their the point of origin of the intermuscular is de-
specimen, and above the epineural, their batable; the bone is not yet ossified in Olney
specimen contains an array of collagen fibers et al.’s larva, and in our dried skeletons it
that does not insert on the epineural, and so might be taken to originate either on the base
is more like the corresponding array in Po- of the neural arch or on the centrum.
lymixia. The other obvious difference is
merely an artifact resulting from G & B hav-
BERYCIFORMES
ing cut away most of Baudelot’s ligament,
the remnant of which is visible as a dark HOLOCENTRIDAE: In their specimen of Hol-
structure at the anteroventral corner of the ocentrus rufus, G & B (their fig. 5) found
myoseptum. We also note here that the first “some peculiar changes of the intermuscular
epineural inserts in a socket at the base of bones in myosepta 6 to 1, which we did not
the neural arch of V1 that is essentially iden- observe in other teleosts.’ Like us (P & J,
tical to that in which the first intermuscular p. 34, Holocentrus (Sargocentron) diadema;
bone of Polymixia inserts. table 7, H. (S.) spinifer, H. (Adioryx) vexil-
In myoseptum 2 of Velifer (fig. 6B) the larius) they found that the intermuscular
epicentral ligament is slender and originates bones of V1 and V2 lie in the horizontal sep-
on the centrum. Above the epineural an array tum. (We placed the origin of both on the
of collagen fibers inserts on the epineural dis- neural arch whereas they placed the origin of
tally, as in myoseptum 1. G & B’s illustration that of V2 on the centrum; cf. Rosen, 1985:
of myoseptum 4 in Velifer (G & B, fig. 3C) fig. 20; P & J: fig. 13.) They interpreted both
shows an array of fibers above the epineural these bones as epicentrals. In the myosepta
similar to that in myoseptum 1 of their spec- of V3-5 G & B found intermuscular bones
imen, but its origin is a bit more concentrated that “‘undergo a remarkable ventral shift” (p.
and it is more extensive, converging on the 22). According to G & B, they are unat-
epineural bone distally and extending beyond tached proximally, and ‘‘are in a position nei-
it. ther epineural nor epicentral but somewhat
G & B did not comment on the POTs (pos- intermediate. There is still an epicentral ten-
terior oblique ligaments) of veliferids. In Ve- don [ligament] beneath these bones. The
lifer we (P & J: 40) recorded the first (most strong epineural tendon [ligament] of the
8 AMERICAN MUSEUM NOVITATES NO. 3312
midbody myosepta is not present here’’ (p. bones of V6 and 7 originate on the rib; un-
12). Instead of the epineural ligament, above like them, we found the bones of V3-—5 also
these bones in myosepta 3-5 G & B illus- attaching directly to the rib (rather than float-
trated a fanlike array of collagen bundles, ing free, as in their H. rufus). If these bones
whose origin on the neural arch or spine is attach to the rib, at least their proximal part
marked with an arrow in G & B’s figure 5. must lie in or penetrate the horizontal sep-
In myoseptum 6 (G & B: fig. 5D) this fanlike tum, which has to be dorsal to the ribs. Here,
array is also present above the intermuscular then, we have unambiguous evidence that an
bone (marked with an asterisk), which orig- epineural bone (G & B acknowledged that at
inates on the rib and which G & B acknowl- least those of V6 and 7 are epineurals) may
edged is an epineural, like the bones and lig- migrate ventrally and enter the horizontal
aments on succeeding vertebrae. The fanlike septum. We (P & J: tables 1-5) found no
array is more extensive in myoseptum 6 than lower teleost in which epineurals entered the
in 5 (G & B: fig. SE), whereas in myoseptum horizontal septum, although we found sev-
4 (G & B: fig. 4F) the fibers are shorter but eral examples (P & J: p. 12) of epineural
more concentrated than in 5. The array be- origin shifted ventrally to the centrum, and
comes a diffuse mass in myosepta 1 and 2 even to the parapophysis (in the aulopiform
(G & B: fig. 5G, H). G & B (p. 22) wrote Scopelosaurus). In holocentrids, our interpre-
of these arrays “‘It is not clear whether an tation of continuity between the intermus-
epineural tendon [ligament] is present as culars of V1 and 2 (in the horizontal septum,
well.”’ They were also “‘not sure how to in- the epicentral position) and those of succeed-
terpret”’ the bones lying between the epineu- ing vertebrae (successively further above that
ral and epicentral position [on V3-—5]. Be- position) is corroborated, in our view, by the
cause we have not seen a similar arrange- occurrence of superficial plates of cartilage
ment in any other species, we are inclined to surrounding the tips of the first three bones
interpret these as autapomorphic for Holo- in our A. diadema (P & J: fig. 13) and the
centridae. Our own interpretation of those first four in H. spinifer (P & J: table 7).
bones is that they are unquestionable epineu- These superficial cartilages may occur at the
rals, because beneath some of them (as G & tips of the first five bones (USNM 304760,
B have shown) there are epicentral liga- three c & s H. (Sargocentron) sp., Aldabra).
ments, and because in other holocentrids they We have rechecked our observations of
attach to the axial skeleton (to the rib; be- holocentrids in c & s Holocentrus ascension-
low). The arrays of collagen fibers above the is (USNM 336175, SL 84 mm) and Plectry-
bones therefore cannot be epineural liga- pops lima (USNM 288860, SL 61 mm), both
ments, and, like the similar arrays above epi- transferred from glycerin to alcohol. In A.
neural bones in Polymixia and Velifer ascensionis, as with the other holocentrids
(above), they are further evidence that simi- that we recorded (P & J: p. 34 and table 7),
lar arrays in other acanthomorphs are not we found the first recognizable epicentral lig-
necessarily epineurals. ament originating on the parapophysis of V9.
Our observations of holocentrids (P & J, It is smaller than that of V10, and consists
p. 34, principally H. (Sargocentron) diade- of two separate, parallel strands, the lower
ma) differed from those of G & B in that we slightly larger than the upper. In the next four
found no epicentral ligaments anterior to V9, myosepta (of V5—8), where G & B (figs. 5B—
whereas in H. rufus they illustrated such lig- E) illustrated epicentral ligaments, we see
aments (their fig. 5) on V4—8 and reported nothing worth calling a ligament, although in
one on V3. We also found that the first POT some of these septa there are much less con-
originates on V13 and inserts on the integu- spicuous double bands of connective tissue
ment together with the epicentral ligament of like that on V9. We infer that these double
V9 (the first that we observed). G & B bands border the junction between the hori-
seemed to confirm this, since in their illus- zontal septum and the myoseptum, where
trations the only POT shown lies about half- there will inevitably be some concentration
way along the epicentral ligament of V10. of connective tissue. In myoseptum 4 there
Like G & B, we found that the intermuscular is a more recognizable epicentral ligament.
2001 JOHNSON AND PATTERSON: ACANTHOMORPH INTERMUSCULARS 9
As in G & B’s illustration of that septum (fig. lustrated what appears to be the same pattern
5F) the ligament runs just below and parallel in A. rufus in their figure 5E (myoseptum 5),
to the intermuscular bone; distally it con- where bone and ligament converge and meet
verges on the bone and attaches to its tip. In distally; in their fig. 5D and F (myosepta 6
myoseptum 3 we see no trace of a ligament. and 4) bone and ligament converge distally,
Nevertheless, if G & B are correct in record- and their tips may be cut off.
ing epicentral ligaments forward to myosep- G & B supported their belief that the in-
tum 3 in Holocentrus, those ligaments have termuscular bones in myosepta 3—8 are not
no bearing on general questions of homology in series with the bones of V1 and 2 (in the
except to reinforce interpretation of the in- horizontal septum) by observation of holo-
termuscular bones in myosepta 3—5 as epi- centrid larvae in which the bones of V1 and
neurals. In their H. rufus, G & B found these 2 are ossified at 11 mm, whereas those of
bones to be unattached proximally. In our H. V3-—6 do not ossify before 26 mm. This dif-
ascensionis the intermuscular bone of V3 is ference (confirmed by our observations of
attached directly to the rib, but those of V4 holocentrid larvae) may be explained by dif-
and 5 are unattached. A very strong sheet of ferentiation of the bone on V2. It is the main
connective tissue originates medial to these site of insertion of the strong sheet of con-
two bones, on the heads of the second and nective tissue described above, is larger than
third ribs (on V4 and 5). It is composed of its neighbors, and has an enlarged head (P &
numerous parallel bands of collagen that run J: fig. 13; Rosen, 1985: fig. 20, H. rufus). In
anterolaterally and insert mainly on the in- any case, delay between ossification of the
termuscular bone of V2, and partially on that intermusculars of Vl—2 and those of succeed-
of V1. As in our other holocentrids, the first ing vertebrae is not evidence that they belong
POT in A. ascensionis originates on the he- to different series. Potthoff and Tellock’s
mal spine of V13 and inserts at the tip of the (1993) account of development in Centro-
epicentral ligament of V9, the first conspic- pomus undecimalis shows that the intermus-
uous one. culars of VI—2 ossify at 7-8 mm, whereas
In Plectrypops lima there are 12 or 13 in- those of succeeding vertebrae do not appear
termuscular bones. The first two originate on until about 13 mm, when the animal has al-
the neural arch. The first is slender; the sec- most doubled in length. In Centropomus
ond is much stouter, with a proximal knob there is no question about serial homology of
fitting a socket at the base of the neural arch. the intermusculars.
Intermusculars 3—8 originate on the rib, 3—4 ANOMALOPIDAE: G & B did not study an-
and 7-8 near the head of the rib, 5—6 further omalopids. In Anomalops (P & J: table 7) we
down the shaft. These intermusculars, 6—7 in recorded a situation much like that in holo-
particular, remain close to the rib for the centrids, with intermuscular bones originat-
proximal part of their course, and the tips of ing on the neural arch and lying in the hor-
1-7 are all on or very close to the lateral line izontal septum on V1 and 2, originating on
nerve. Intermusculars 9—11 originate succes- the rib on V3 and 4, on the parapophysis on
sively higher on the hemal spine, 12 on the V5-10, the centrum on V11, and thereafter
centrum, and 13 (which is diminutive) on the a series of epineural ligaments ascending
neural arch. The tips of intermusculars 8—12 from the centrum to the neural arch in suc-
end successively further above the lateral line cessive segments. We found the bones of
nerve. In the epicentral series, we find rec- V3-6 to be partially in the horizontal sep-
ognizable ligaments forward to V5, where tum. Our first epicentral ligament was on
the ligament originates on the rib, distal to V11, beneath the last ossified intermuscular.
the origin of the intermuscular bone. The epi- We found the first POT originating also on
central ligaments of V5—7 run out close be- V11, and wrote (P & J: p. 40) “Because
low the bone, and converge with it distally there are no epicentrals anterior to this, POTs
to attach to its tip (as with the ligament of of V11-—13 lie free in the horizontal septum
V4 in H. ascensionis, above). Beyond V7, as they pass forward after crossing the epi-
bone and ligament diverge distally, succes- central of V11.’’ Checking this and two other
sively more in each myoseptum. G & B il- c & S specimens after transfer to alcohol, we
10 AMERICAN MUSEUM NOVITATES NO. 3312
found discrete epicentral ligaments extending specimen is a poor one) we again found a
forward to V9 and, additionally, one attach- situation like that in holocentrids: bones on
ing to the epineural bone of V1 and some- the neural arch and in the horizontal septum
times V2. More diffuse collagen arrays at- on V1 and 2, partially in the horizontal sep-
tach to the ribs of V3—V8. In one specimen tum on V3—6, where they originate on the
the last epineural bone is on the neural arch rib (3—5) or parapophysis (6), and ascending
of V14. Posteroventrally directed fiber arrays thereafter to the centrum and eventually (as
(similar to those identified as epineurals by ligaments) to the neural arch. We found the
G & B in Hoplostethus, see below) are de- first epicentral ligament on V9. The epineural
tectable from V1 to V8-9, the posteriormost arrangement is much the same in a specimen
the weakest. Dissection of a whole alcoholic of Beryx splendens (USNM 306134, SL 116
specimen revealed the first seven intermus- mm), where the epicentral ligaments extend
culars partially in the horizontal septum. forward to at least V8.
In Photoblepharon (P & J: table 7) we re- TRACHICHTHYIDAE: We have checked our
corded a pattern different from that in An- observations of Hoplostethus (P & J: table 7)
omalops, with a gap in the epineural series in three new (and much better) c & s speci-
between the two bones originating on the mens of H. mediterraneus transferred to al-
neural arch on V1 and 2 and a ligament on cohol. On the bases of the hemal arches of
the rib of V6. Thereafter we recorded an epi- V8-—13 there are conspicuous ligaments that
neural ligament on the rib of V7, ligaments we recorded (on V9-—13) as epicentrals. We
with included bone on the parapophyses of were wrong. G & B correctly interpreted
V8-—12, and ligaments on the neural arch or these as epineurals, and believe that they, to-
spine beyond that. We placed the ligaments gether with the horizontal septum and its in-
of V6-8 partially in the horizontal septum. cluded epicentrals, underwent ‘‘a_ ventral
Our first epicentral ligament was on the par- shift’’ in the middle part of the body, from
apophysis of V10. Checking these observa- about V6—15. We (P & J: table 7) recorded
tions in two larger (69 and 80 mm) c & s intermuscular bones (our epineurals) on V1
specimens of Photoblepharon that were and 2 in Hoplostethus, but found no epineu-
transferred from glycerin to alcohol, we ral ligaments beyond them. G & B (fig. 4)
found that the first ligament originates on the illustrated what they called epineural liga-
rib of V4, and that it and the succeeding ones ments in myosepta 6, 3, and 1; these are dif-
are entirely in the horizontal septum and so fuse arrays of collagen bundles, similar to
in the epicentral position. They are relatively those they illustrated on V1 in Polymixia
broad and could be epineurals, epicentrals, or (compare their figs. 2 and 4D, of myoseptum
a coalescence of the two, because on V8 and 1 in Polymixia and Hoplostethus) and on V1
V9 epicentral ligaments and epineural bones and 4 in Velifer (compare their figs. 3C and
(or ligaments with included bone) share the AC, of myoseptum 4 in Velifer and 3 in Ho-
same origin, at the junction of rib and para- Plostethus). On V5-—7 in Hoplostethus these
pophysis, and diverge distally. Beyond that, arrays originate broadly low on the neural
epineural ligaments (with included bone dis- arch, whereas in more anterior vertebrae their
tally to V12) gradually ascend the centra, the line of attachment ascends the arch, until on
last inserting on the base of the neural arch V1 it is at the base of the neural spine Gn G
of V15—17, while the epicentral ligaments & B’s fig. 4D, of myoseptum 1; notice the
continue posteriorly on the parapohyses. As distance between it and the intermuscular
in Anomalops, there are fiber arrays anteri- bone, inserting on the neural arch). There-
orly (similar to those identified as epineurals fore, these arrays reverse the normal pattern
by G & B in Hoplostethus, see below), the for epineurals, which always move from the
last and weakest one above the epineural neural arch anteriorly to the neural spine pos-
bone of V8. The first POT in Photoblephar- teriorly. Further, the fibers are directed pos-
on originates on V14 and inserts at the tip of teroventrally, rather than posterodorsally, the
the ligament of V9. normal orientation of epineurals. In two of
BERYCIDAE: G & B did not study berycids. our specimens (USNM 307556, SL 73 mm;
In Centroberyx affinis (P & J: table 7; our USNM 307273, SL 75 mm) the array of fi-