UN
CO
RR
EC
TE
D P
RO
OF
When two equals three: developmental osteology and homology of the
caudal skeleton in carangid fishes (Perciformes: Carangidae)
Eric J. Hiltona,
 and G. David Johnsonb
aDivision of Fishes, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560, USA
bDivision of Fishes, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560, USA

Author for correspondence (email: ehilton@fieldmuseum.org)
SUMMARY Ontogeny often provides the most compelling
evidence for primary homology in evolutionary developmental
studies, and is critical to interpreting complex structures in a
phylogenetic context. As an example of this, we document the
ontogenetic development of the caudal skeleton of Caranx
crysos by examining a series of cleared and stained larval and
post-larval specimens. By studying ontogeny, we are able to
more accurately identify some elements of the adult caudal
skeleton than is possible when studying the adult stage alone.
The presence of two epurals has been used as a
synapomorphy of Caranginae (homoplastically present in the
scomberoidine genera Scomberoides and Oligoplites). Here
we find that three epurals (ep) are present in larvae and small
post-larval juveniles (i.e.,o25mm standard length [SL]) of
C. crysos and other carangines, but ep2 never ossifies and
does not develop beyond its initial presence. Ep2 was last
observed in a 33.6mm SL specimen as a small nodule of very
lightly stained cartilage cells and eventually disappears
completely. Therefore, the two epurals present in the adult
are ep1 and ep3. In other carangines examined (e.g., Selene,
Selar) the rudimentary ep2 ossifies and appears to fuse to the
proximal tip of ep1. In these taxa, therefore, the two epurals of
the adult appear to be ep112 and ep3.We found no indication
of three epurals at any stage in the development of Oligoplites
(developmental material of Scomberoides was unavailable).
We discuss the osteology of the caudal skeleton of carangoid
fishes generally and emphasize the power and importance of
ontogeny in the identification of primary homology.
Embryology affords further a test for homologies in contradis-
tinction of analogies.
(Louis Agassiz, Essay on Classi?cation, 1857, p. 86.)
INTRODUCTION
The importance of ontogeny (Agassiz?s ??embryology??) for
assessing primary homology (sensu de Pinna 1991) has long
been recognized (e.g., Huxley 1859) and was recently under-
scored as a largely untapped source for data in phylogenetic
analyses of ?shes (Johnson 1993; also see Leis et al. 1997 and
articles in Moser et al. 1984). Although ontogeny is only one
of several lines of evidence for distinguishing homologies from
analogies (others include topographic correspondence and the
principle of connectivity; Rieppel and Kearney 2002) it may
be one of the most powerful and is certainly one of the most
understudied or underutilized. The sequence of ontogenetic
events may be useful as phylogenetic character data, although
perhaps more significant is the role of ontogeny in identifying
the components of a morphological complex. Simply stated,
the study of the adult condition alone may be inadequate to
conclude that similar conditions re?ect common ancestry or
vice versa.
The complexity and broad diversity of the skeleton of
?shes makes them ideal for the study of ontogeny in relation
to phylogeny and functional systems. At the root of under-
standing this complexity, however, is the accurate establish-
ment of primary homology statements for comparisons. The
caudal skeletons of adult teleostean ?shes vary widely in their
composition and overall anatomy (Whitehouse 1910; Monod
1968; Fujita 1990) and range from taxa in which all elements
of the complex remain separate from one another (e.g., Hio-
don, Hilton 2002) to those in which nearly all elements fuse
together (e.g., Luvarus, Tyler et al. 1989). With the consoli-
dation of individual parts, either through ontogeny or phy-
logeny, their identi?cation and comparison between taxa may
be dif?cult by studying adult anatomy alone (see Arratia and
Schultze 1992). One extreme case of ontogenetic fusion is
found in the caudal skeleton of scombroid ?shes (tunas and
their allies), in which hypurals 1?4 fuse with the urostyle into
a single ??hypural fan?? (Collette and Chao 1975; Pottho?
1975; Pottho? et al. 1986). In the tuna Thunnus atlanticus, as
E D E 1 4 8 B Dispatch: 11.2.07 Journal: EDE CE: SangeethaJournal Name Manuscript No. Author Received: No. of pages: 12 PE: Suseela/Suresh
EVOLUTION & DEVELOPMENT 9:2, 178?189 (2007)
& 2007 The Author(s)
Journal compilation & 2007 Blackwell Publishing Ltd.
178
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OF
described in detail by Pottho? (1975), the neural arch of pre-
ural centrum 2, and epural 1 are also incorporated into the
hypural fan, thereby forming a very rigid attachment site
for the ligaments and musculature associated with strong
swimming.
The caudal skeleton of teleostean ?shes is often a rich
source for phylogenetically informative characters and is
among the ?rst character systems to be examined in osteo-
logical studies of ?shes. During our ongoing study of the
osteology of jacks and their relatives (Carangidae), we dis-
covered an interesting example of the impact that compara-
tive ontogenetic studies may have on the interpretation of
phylogenetic patterns in adult morphology. In this article, we
document the ontogenetic development of the caudal skeleton
of the blue runner, Caranx crysos, based on a series of cleared
and stained larval and postlarval specimens as well as dry
skeletons of adults, and accordingly we are able to precisely
identify some elements of the caudal skeleton of the adult. We
also present observations on the development of the caudal
skeleton in certain other related taxa and discuss the impor-
tance of ontogeny in the study of comparative anatomy.
C. crysos is a member of the family Carangidae, the jacks,
pompanos, and trevallies. Carangidae are part of the Cara-
ngoidei (Fig. 1), a monophyletic group of percomorph ?shes
that also includes the rooster?sh (Nematistiidae), the remoras
(Echeneidae), the cobia (Rachycentridae), and the dolphin-
?shes (Coryphaenidae). As popular game and food ?shes, this
is an economically important group of ?shes, many of which
super?cially resemble scombroid ?shes (e.g., tunas and mack-
erels). These ?shes have been studied from functional and
physiological standpoints (e.g., Grubich 2003) and are the
nominal model for carangiform locomotion (Breder 1926;
Gray 1968). Although there is much information available on
the skeletal anatomy of carangoid ?shes (e.g., Starks 1911;
Suzuki 1962; Smith-Vaniz 1984), their developmental oste-
ology remains poorly known. Few studies have dealt with
aspects of the development of the skeleton in carangids, and
these have been at varying levels of detail and mostly made
before the use of alcian staining of cartilage had come into
practice (e.g., Schnakenbeck 1931; Ahlstrom and Ball 1954;
Aprieto 1974; Miller and Sumida 1974; Laroche et al. 1984;
Liu 2001). External aspects of larval development of C. crysos
have been relatively well studied (e.g., McKenney et al. 1958;
Berry 1959), but the ontogeny of its skeleton has not.
C. crysos was chosen for the developmental component of our
study because it is abundant and all life history stages could
be well represented in our sample; for many carangid taxa this
is not the case. With this article, we hope to reiterate the
importance of incorporating ontogenetic information into
comparative morphological studies.
MATERIALS AND METHODS
Preparation, study, and illustration
Most specimens examined were cleared and stained for bone and
cartilage (e.g., Dingerkus and Uhler 1977); some specimens less
than 12mm standard length (SL) were prepared using an alcohol?
alizarin solution as recommended by Springer and Johnson (2000).
Dry skeletons of adult specimens were prepared by the method of
Bemis et al. (2004). Specimens were examined under a binocular
Zeiss Stemi SV 11 dissecting microscope with a camera lucida and
substage illumination. Illustrations were rendered using Adobe Il-
lustrator software based on camera lucida sketches or digital im-
ages taken with a Nikon COOLPIX 8700 coupled to the dissecting
microscope.
Specimens examined
Cleared and stained (c&s) and dry skeletal (ds) specimens were
examined, representing pre- and post-?exion larvae, juveniles, and
adults ranging from 2.7mm notochord length (NL) to 329mm SL.
The following specimens of C. crysos were examined: MCZ 64775
(2 c&s, 9.7?11.9mm SL); MCZ 84131 (8 c&s, 5.7?24.7mm SL);
MCZ 84135 (2 c&s, 6.0?7.6mm SL); SEAMAP st-568 (1 c&s,
13.1mm SL); SEAMAP s-574 (23 c&s, 2.7mm NL?5.3mm SL);
SEAMAP 1934 (5 c&s, 3.5mm NL?4.7mm SL); SEAMAP 17731
(1 c&s, 18.2mm SL); SEAMAP 17747 (4 c&s, 5.9?7.1mm SL);
SEAMAP 17792 (2 c&s, 6.5?8.9mm SL); SEAMAP 17788 (5 c&s,
6.2?9.6mm SL); SEAMAP 18397 (15 c&s, 2.3mm NL?4.8mm
SL); SEAMAP 18504 (5 c&s, 4.4?14.9mm SL); SEAMAP 18513
(1 c&s, 11.2mm SL); SEAMAP 18659 (2 c&s, 5.7?10.6mm SL);
UMA F11174 (1 ds, 185mm SL); UMA F11173 (1 ds, 220mm
SL); UMA F11596 (1 ds, 250mm SL); UMA F10696 (1 ds,
329mm SL); USNM 158836 (13 c&s, 10.9?100mm SL).
Cleared and stained specimens of taxa representing all other
carangid genera were also studied, and specimens are cited in the
text when relevant to the discussion. This material includes devel-
opmental series of the scomberoidine Oligoplites saurus (approxi-
mately 40 specimens from 3.8mm SL to adults) and the carangine
Fig. 1. Hypothesis of the relationships of the major groups of
Carangoidei. Relationships of carangid subfamilies following
Smith-Vaniz (1984), Gushiken (1988), and Reed et al. (2002).
Those of noncarangid families following Johnson (1984) and
Smith-Vaniz (1984).
Q1
Q2
Developmental osteology of carangid ?shes 179Hilton and Johnson
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OF
Selene vomer (approximately 30 specimens from 4.1mm SL to large
adults); both of these series are largely derived from the SEAMAP
collection.
Anatomical abbreviations
dcr, distal caudal radial; ep, epural; hpu, hypurapophysis; hspu,
hemal spine of pu; hy, hypural; k, dorsal and ventral ?eshy keels on
the lateral surface of the caudal ?n; lppu4, lateral process on pu4;
nc, notochord; nspu, neural spine of pu; phy, parhypural; pu, pre-
ural centrum; sc, scutes; uc, ural centrum; ust, urostylar centrum
(5pu11uc).
Institutional abbreviations
MCZ, Museum of Comparative Zoology, Harvard Universi-
ty, Cambridge, MA; SEAMAP, Southeast Area Monitoring
and Assessment Program Ichthyoplankton Archiving Center,
Fish and Wildlife Research Institute, St. Petersburg, FL;
UMA, University of Massachusetts Amherst Zoological
Collections, Amherst, MA; USNM, United States National
Fig. 2. Adult caudal skeleton of
Caranx crysos. (A) External and
(B) internal views showing posi-
tion of scutes and ?n rays. (C)
Dorsal, (D) lateral, and (E) ventral
views of internal caudal skeleton.
Scales are omitted from A. Ante-
rior facing left. (A, B) UMA
F11174, 185mm standard length
(SL); (C?E) UMA F11596,
250mm SL.
180 EVOLUTION & DEVELOPMENT Vol. 9, No. 2, March^April 2007
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OF
Museum of Natural History, Smithsonian Institution, Wash-
ington, DC.
RESULTS
General anatomy of the caudal peduncle and fin
supports
Berry (1960) described and illustrated the development of the
scutes and scales of C. crysos, including those of the caudal
peduncle, and their development will not be described here as
our ?ndings generally agree with his. We do provide a brief
description of the general adult anatomy of this region as a
reference for our study of the caudal skeleton. The posterior
portion of the adult caudal region is illustrated in Fig. 2.
Externally, it bears a series of thick overlapping scutes, which
are particularly deep and robust in the region posterior to pu4
and form a well-developed lateral keel on the caudal peduncle
(Fig. 2A). Posteriorly, the scutes decrease in size, and the
series terminates at about the level of the posterior margin of
the hypurals. Extending onto the caudal ?n are ?eshy but
rigid lateral keels dorsal and ventral to the scutes, which
greatly sti?en the central portion of the deeply forked caudal
?n. The dorsal and ventral caudal ?n rays deeply embrace the
endoskeletal supports (Fig. 2B), and small unsegmented pro-
current ?n rays extend the caudal ?n anteriorly to about the
middle of the centrum of pu2. The skeletal supports of the
caudal ?n of C. crysos, like those of other carangids, involve
vertebrae posterior to and including pu3.
Vertebral centra
Ossi?cation of centra progresses rostrocaudally. By 3.5mm
NL the centra of the immediate preural region are ossi?ed as
very thin, smooth rings of bone. The compound urostylar
centrum (5pu11uc) is bent dorsally, as it follows the curve
of the notochord (Fig. 3, B?D) and ossi?es as a single element
by 4.7mm SL. In later stages, this centrum is fused to the
uroneural and hypural 5 (Figs. 2D and 4D). Lateral struts on
the centra are evident by 25.2mm SL (e.g., Fig. 4D). A
prominent lateral process on pu4 (Fig. 2, C and E) is well
developed by 25.2mm SL.
Fig. 3. Caudal skeleton of Caranx
crysos in early stages of develop-
ment. Anterior facing left. Cartil-
age is shown in gray, bone in
stipple; white stipple indicates light
ossi?cation. A, specimen from
SEAMAP 18504; (B?D) specimens
from MCZ 84131.
Developmental osteology of carangid ?shes 181Hilton and Johnson
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OF
Parhypural and hypural series
The hypochordal elements of the caudal skeleton include the
parhypural (phy) and ?ve hypurals (hy1?5). The smallest
specimens available did not stain well (not illustrated) but the
phy, hy1, and hy2 are present by 3.3mm NL; hy3 and hy4 are
also present by 3.8mm NL (at this stage, hy4 is not well
developed, although its proximal margin is well de?ned). The
phy, hy1, and hy2 are fused proximally early in ontogeny
(Fig. 3A): in a 4.3mm NL specimen, hy1 and hy2 are fused
and the phy is fused to hy112 proximally although it remains
separate for most of its length. The smallest well-stained
specimen (4.6mm NL) is an early post-?exion larva. Al-
though the elements are continuous proximally in this and
other specimens of a comparable stage, there are distinct sep-
arations between the phy and hy1, and hy1 and hy2. In later
stages (by 4.4mm SL; Fig. 3A), hy1 and hy2 are fused distally
and, when hy1 and hy2 become more completely formed,
there is a distinct fenestra marking the boundary between the
two hypurals. This large fenestra is present in the proximal
portion of hy112 (e.g., Figs. 3 and 4) and becomes progres-
sively smaller as the hypurals ossify along its distal margin
(Fig. 5, A?C). It is still present in a 19.5mm SL specimen but
disappears by 25.2mm SL (in Fig. 4, B and C, the fenestra is
hidden by the hypurapophysis). Hypural 3 and hy4 are com-
pletely separate in a 9.7mm SL specimen but are fused distally
in a 10.9mm SL specimen; they are completely fused by
25.2mm SL.
Hypural 5 is the smallest of the series and is the last to
develop. It was ?rst seen in 5.7mm SL larvae (MCZ 84131,
SEAMAP 18659) as a small cartilage positioned along the
dorsal margin of hy4 (Fig. 3B); ossi?cation was ?rst detected
in a 7.2mm SL specimen. Hypural 5 gradually elongates and
becomes positioned between the dorsal margin of hy4 and the
urostyle (e.g., Fig. 3D) and remains autogenous until at least
15.5mm SL, but eventually fuses to the urostyle (e.g., Fig.
4D). In the adult, there is a groove indicating the boundary
between the two elements, but they cannot be separated and
the groove is indistinct proximally.
Epurals (ep)
Three epurals (ep1-3) are present in larvae and small
(i.e.,o25mm SL) post-larval juveniles (Figs. 3?5). The ep ap-
pear anterior to posterior and in the earliest stages decrease in
size anterior to posterior (Fig. 3A). In slightly later stages ep1
and ep3 become the largest, with ep2 being about half their
length (e.g., Fig. 3B). All three epurals are present by 4.2mm
NL. Epural 1 bears an anterior projection that is recognizable
early in ontogeny (Fig. 3B) and continues to grow anteriorly
so that it becomes intercalated between the rudimentary
neural spine of pu2 (nspu2) and nspu3 (Fig. 4). Epural 2 does
Fig. 4. Caudal skeleton of Caranx
crysos in various stages of larval
and post-larval development in lat-
eral view. Anterior facing left. Car-
tilage is shown in gray, bone in
stipple; white stipple indicates light
ossi?cation. (A?D) specimens
from USNM 158836.
182 EVOLUTION & DEVELOPMENT Vol. 9, No. 2, March^April 2007
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OF
not appear to develop further after its initial appearance and
eventually disappears. It was last detected in a 33.6mm SL
specimen (USNM 158836) as a small nodule of very lightly
stained cartilage cells positioned between the proximal por-
tions of ep1 and ep3; in a 40.3mm SL specimen from the
same lot there was no space between ep1 and ep3 and no trace
of ep2.
Uroneural
The paired dermal uroneurals are the last elements of the
caudal skeleton to develop and were ?rst found in a 6.6mm
SL specimen as thin splints of bone lying dorsal and lateral to
the notochord posterior to ep3 (e.g., Fig. 3C). These splints
grow anteriorly, ventral to the epural (ventral to ep2 by
7.0mm SL and ventral to ep1 by 8.0mm SL). The uroneurals
were ?rst found to be fused posteriorly to the urostylar cen-
trum at 14.9mm SL (Fig. 4B) but become fused to the uro-
stylar centrum along their entire ventral margins in later
stages (e.g., Fig. 4, C and D). In an 8.9mm SL specimen
(MCZ 84131) a second more anterior uroneural was found
lying along the proximal tips of the epurals on both sides of
the specimen (Fig. 3D). This specimen is considered anom-
alous, as an autogenous second pair of uroneurals was not
observed in any other specimen. It is possible that the uro-
neural of later stages is actually best regarded as un112 (as in
Coryphaena, Pottho? 1975), but this hypothesis requires more
data.
Fig. 5. Caudal skeleton of carangid ?shes at various stages of development. (A?C) Caranx crysos (MCZ 84131, SEAMAP 18504, and
USNM 158836, respectively). (D?F) Selene vomer (SEAMAP 26870, SEAMAP 26881, and USNM 383080, respectively). (G?I) Oligoplites
saurus (SEAMAP 29375, SEAMAP 16618, and SEAMAP 17031, respectively). Anterior facing left.
Developmental osteology of carangid ?shes 183Hilton and Johnson
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OF
Neural and hemal elements of preural skeleton
The neural and hemal spines of pu2 and pu3 are distinct from
more anterior ones. Early in ontogeny they are very thick carti-
laginous structures, whereas themore anterior neural and hemal
spines are much more slender (e.g., Fig. 3B). On the vertebrae
that support the caudal ?n, the neural and hemal arches ossify
separately from their respective spines (e.g., Fig. 3, B and C)
but become continuous soon after ossi?cation (e.g., Fig. 3D).
The nspu2 is always rudimentary even at ?rst development
(e.g., Fig. 3A) and disappears in the adult (e.g., Fig. 2D).
Distal caudal radials
The elongate distal caudal radial (5post-hypural cartilage
and post-hemal spine cartilages of Fujita 1990) associated
with hspu3 is the largest; it was ?rst detected in a 5.7mm SL
specimen (e.g., SEAMAP 18659; although see Fig. 3B), and
it is well developed by 6.2mm SL (SEAMAP 17788). Two
additional spherical cartilages are associated with the distal
tips of hy5 and hspu2 and were ?rst found in a 6.6mm SL
specimen (SEAMAP 17788).
DISCUSSION
Variation of the caudal skeleton of Carangoidei
The adult caudal skeleton of carangoid ?shes shows a wide
variety of patterns. Rosenblatt and Bell (1976: ?gs. 10, 11)
?gured the caudal skeleton of Nematistius. In the adult (rep-
resented in their ?gures by a large juvenile) there are three
independent epurals (ep2 and ep3 are equal in length), two
uroneurals (both of which remain autogenous from pu11uc),
a parhypural, and three hypurals (their hy112, hy314, hy5).
Pottho? (1980: 298) described the caudal complex of adult
Coryphaena as comprising ??two hemal spines, a parhypural, a
ventral, and dorsal hypural plate [each formed by ontogenetic
fusion of two hypurals], hypural 5, a uroneural pair (fused
from two pairs), and an epural (fused from two),?? the neural
spines of pu2 and pu3, and the three associated centra (pu2,
pu3, and ust).
Rachycentron and echeneids have ?ve independent hypur-
als, whereas Coryphaena and carangids have three distinct
hypurals, representing hy112, hy314, and hy5 in taxa for
which the ontogeny has been studied (e.g., Pottho? 1980;
Fujita 1990, also personal observation). In carangids, hy5 be-
comes tightly enveloped by the uroneural dorsally, and the
posterior extension of the urostylar centrum laterally. Hy-
pural 5 remains autogenous in most carangids but may be-
come fused to the compound urostylar centrum-uroneural.
The taxonomic distribution of this fusion needs further study,
but it was not observed in noncarangine carangids (e.g.,
Naucrates, Elagatis, Trachinotus, Oligoplites; personal obser-
vation, also taxa illustrated by Hollister, Monod, Fujita, and
others). Similarly, fusion between the uroneural and the uro-
stylar centrum was only observed in carangines.
There have been many previous descriptions and illustra-
tions of the caudal skeleton of carangids [Scomberoidinae:
Scomberoides and Parona (Smith-Vaniz and Staiger 1973);
Trachinotinae: Trachinotus (Hollister 1941; Monod 1968;
Fujita 1990); Naucratinae: Seriola and Elagatus (Berry 1969;
Fujita 1990; Liu 2001); Caranginae: Decapterus, Alectis, Sel-
ene, Trachurus, Caranx, Selar, and Chloroscombrus (Hollister
1941; Monod 1968; Fujita 1990; Suda 1996)]. Within the
family Carangidae the caudal skeleton is remarkably con-
stant, although there is subtle variation in various fusion pat-
terns and number of elements in the subfamily Caranginae
(Table 1).
Loss versus fusion of epurals
Perhaps the most interesting variation in the carangid caudal
skeleton that we discovered during this study is in the number
of epurals (Table 1). Nematistius has three epurals (Rosenblatt
and Bell 1976), as do echeneids and Rachycentron (e.g., Fujita
1990, also personal observation.). Pottho? (1980) demon-
strated that the single epural of Coryphaena is the product of
fusion of two independent epurals during ontogeny. In all
noncarangine genera (i.e., Trachinotinae, Scomberoidinae,
and Naucratinae) except for the scomberoids Scomberoides
and Oligoplites, there are three epurals (Smith-Vaniz 1984:
table 127). Based on the current understanding of the inter-
Table 1. Some characters in the caudal skeletal anatomy
of carangoid ?shes
Taxon1
Number of
epurals (in adult)
Unfused
to ust
hy5 fused
to ust
Nematistiidae 3 No No
Coryphaenidae 1 No No
Rachycentridae 3 No No
Echeneidae 3 No No
Carangidae
Scomberoidinae 2 or 3 No No
Trachinotinae 3 No No
Naucratinae 3 No No
Caranginae 2 Yes No or yes
1For ease of use with our discussion, we here list the genera included
in the families and subfamilies of Carangoidei, following Smith-Vaniz
(1984).
Nematistiidae: Nematistius. Coryphaenidae: Coryphaena. Rachycentridae:
Rachycentron. Echeneidae: Echeneis, Phtherichthys, Remora, Remorina.
Carangidae: Scomberoidinae: Oligoplites, Parona, Scomberoides; Trachi-
notinae: Lichia, Trachinotus; Naucratinae: Campogramma, Elagatis, Nau-
crates, Seriola, Seriolina; Caranginae: Alectis, Alepes, Atule, Atropus,
Carangichthys, Carangoides, Caranx, Chloroscombrus, Decapterus, Gnath-
anodon, Hemicaranx, Megalaspis, Pantolabus, Parastromateus, Pseudoc-
aranx, Selar, Selaroides, Selene, Trachurus, Ulua, Uraspis.
ust, urostylar centrum.
184 EVOLUTION & DEVELOPMENT Vol. 9, No. 2, March^April 2007
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OF
relationships of carangoid ?shes (Fig. 1), it can be concluded
that three epurals is plesiomorphic for the family Carangidae.
Adults of all genera of the subfamily Caranginae have only
two epurals, and Smith-Vaniz (1984) considered this to be a
synapomorphy of the group (independently derived in the
scomberoids Scomberoides and Oligoplites). However, he did
not identify which epurals are present in carangines and the
two scomberoidine genera, only that there are two rather than
three.
We have documented here the ontogenetic loss of ep2 in
C. crysos, and thus can identify the two epurals present in
the adult as ep1 and ep3. In examining comparative material,
we discovered that larval and small juvenile specimens of
Selene vomer also have a third epural (ep2) similar in form to
that of small specimens of C. crysos (i.e., a short bar of
cartilage; Figs. 5D and 6A) but ossi?ed (Figs. 5, E, F and
6, B, C). Larger specimens (i.e., in the adult condition) have
only two epurals (Fig. 6D). We observed a similar ossi?ed ep2
in small (i.e., juvenile) specimens of the carangines Selar
crumenophthalmus (MCZ 149687, 30.7?56.4mm SL; Fig. 7),
Chloroscombrus chrysurus (USNM 383076, 18.4?26.7mm SL;
USNM 383077, 54?62mm SL), and Gnathanodon speciosus
(USNM 383087, 86mm SL). Larval T. trachurus have
three epurals (Schnakenbeck 1931) and Witzell (1977:
Fig. 7) ?gured a similar arrangement in a juvenile specimen
of Parastromateus niger, as did Fujita (1990: ?gs. 295, 297
respectively) in Caranx ignobilis and Alectis ciliaris. In
larger specimens of these and other carangine taxa, there
are only two epurals (e.g., Fig. 6D; also see Trachurus
japonicus and Caranx equula, Fujita 1990: ?gs. 294, 296,
respectively).
Further detailed study of the development of the caudal
skeleton in carangines is necessary. Although our limited se-
ries of the carangine Selene vomer does not demonstrate the
actual fusion of ep2 to ep1, it strongly suggests that such
fusion occurs, a condition different from that in C. crysos, in
which ep2 is lost. This hypothesis gains support (albeit
through circumstantial evidence) from our observation of two
specimens of Selar, also a carangine, in which ep1 and ep2 are
fused (MCZ 14968; 36.2mm SL and 42.2mm SL; Fig. 7). In
the 42.2mm SL specimen (Fig. 7C) the two are completely
fused, but an inclusion in the bone of the anteriormost epural
that is the exact shape and position of an autogenous ep2 in
other specimens can be detected with transmitted light. There
is some individual variation in fusion, as the two epurals are
clearly separate in the largest specimen of this lot (56.4mm
SL). In the 70mm specimen of T. trachurus illustrated by
Schnakenbeck (1931: Fig. 3f) ep2 appears to be continuous
with ep1. In taxa in which ep2 ossi?es, it may persist as an
autogenous element in relatively large individuals (e.g., the
Fig. 6. Caudal skeleton of Selene
vomer at four developmental
stages. (A) SEAMAP 26881. (B)
USNM 383080. (C) USNM
383081. (D) UMA F10936. Ante-
rior facing left. Cartilage is shown
in gray, bone in stipple.
Developmental osteology of carangid ?shes 185Hilton and Johnson
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OFFig. 7. Caudal skeleton of Selar crumenophthalmus
(all from MCZ 149687). Dashed line in C indicates
the outline of the inclusion of ep 2 in the compound
ep112. Anterior facing left. Cartilage is shown in
gray, bone in stipple.
Fig. 8. Caudal skeleton of Oligo-
plites saurus in early stages of de-
velopment. (A, B, and C) from
SEAMAP 29375. (D) from SEA-
MAP 16618. Anterior facing left.
Cartilage is shown in gray, bone in
stipple.
186 EVOLUTION & DEVELOPMENT Vol. 9, No. 2, March^April 2007
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OF
150mm SL specimen of Alectis ciliaris illustrated by Fujita
1990: ?g. 297).
The question then arises, ??What is the ontogenetic trajec-
tory in the scomberoidines that also have two epurals in the
adult (i.e., Oligoplites and Scomberoides)??? Larval material of
Scomberoides was not available to us and there is no data on
its development in the literature. In the larval and post-larval
specimens of O. saurus we examined, there is no trace of a
third epurals at any stage (Figs. 8 and 9). The anteriormost
epural is always larger than ep2 but becomes much broader in
juveniles and adults than in larvae. The general pattern of
development was found to be consistent within our series,
with only a few minor variations noted. For example, in a few
specimens ep1 is roughly J-shaped (Fig. 8B), although the
anterior portion of this element was not found to be autog-
enous in either smaller or larger specimens (e.g., Fig. 8, A and
C). A small post-larval specimen of O. saurus (USNM
383088, 13mm SL) has an anomalous condition of a single
epurals, which is likely the fusion of two epurals. We conclude
that the two epurals of the scomberoidine and carangine
genera do not represent similar conditions even though a
similar pattern is found in adults of the two groups. However,
it should be noted that ep1 of Oligolplites is not necessarily
equivalent to that of Caranx or other carangines.
CONCLUSION
Reduction in the number of discrete elements through on-
togeny may occur by fusion, where adjacent elements become
inseparable from one another, or by loss, where an element
initially forms but disappears through resorption. Both mech-
anisms result in a similar con?guration but are fundamentally
different in that they re?ect different ontogenetic trajectories.
Our example from the development of the caudal skeleton of
carangids further demonstrates the importance of ontogeny in
formulating primary homology statements and in character
conceptualization, and the potential fallacy of a priori as-
sumptions about phylogenetic or ontogenetic fusion or loss
(see discussions by Nelson 1969; Patterson 1977; Donoghue
2002; Hilton 2002; and Britz and Johnson 2003, 2005).
Although beyond the scope of this project, a broader
taxonomic comparison of the development of the caudal
skeleton within carangines may reveal additional patterns of
ontogenetic variation, possibly re?ective of phylogenetic pat-
terns. It is clear that the two epurals of adult Scomberoides
and Oligoplites are not entirely comparable with those of
Caranginae. There have been recent advances in the study of
phylogenetic systematics of Carangidae (e.g., Reed et al.
2002). However, a better understanding of the relationships
Fig. 9. Caudal skeleton of Oligo-
plites saurus in various stages of
larval and post-larval develop-
ment. (A, B) from SEAMAP
16618. (C) from USNM 383078.
(D) USNM 383079. Anterior fac-
ing left. Cartilage is shown in gray,
bone in stipple.
Developmental osteology of carangid ?shes 187Hilton and Johnson
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OF
within Carangidae, in particular among the genera in the
subfamily Caranginae (e.g., more complete taxon sampling),
is necessary before the true significance of the variation in
ontogenetic trajectories can be fully assessed. The different
ontogenetic trajectories arriving at similar adult conditions
described herein may also be re?ective of different gene ex-
pression patterns, and as such, carangid ?shes may o?er a
useful model system for both evolutionary and developmental
approaches to understanding anatomy.
Acknowledgments
We thank M. Leiby and K. Williams (SEAMAP) for making avail-
able most of the early stages of the developmental series, indeed most
of the specimens, used in this project. We also thank K. Hartel
(MCZ) and W. E. Bemis (UMA) for access and permission to pre-
pare specimens in their care. W. F. Smith-Vaniz and V. G. Springer
offered encouragement and suggestions with our study of carangid
osteology. R. Britz, V. G. Springer and two anonymous reviewers
offered helpful suggestions on a previous version of this manuscript.
This research was conducted while E. J. H. was supported by a
Smithsonian Institution Postdoctoral Fellowship in the Department
of Zoology, for which he is grateful.
REFERENCES
Agassiz, L. 1857. Contributions to the Natural History of the United States of
America. First Monograph. In Three Parts.?I. Essay on Classi?cation.?II.
North American Testudinata.?III. Embryol ogy of the Turtle; with Thirty-
Four Plates. 1. Little, Brown and Company, Boston.
Ahlstrom, E. H., and Ball, O. P. 1954. Description of eggs and larvae of
jack mackerel (Trachurus symmetricus) and distribution and abundance
of larvae in 1950 and 1951. Fish. Bull. 56: 207?245.
Aprieto, V. L. 1974. Early development of ?ve carangid ?shes of the Gulf
of Mexico and the south Atlantic coast of United States. Fish. Bull. 72:
415?443.
Arratia, G., and Schultze, H.-P. 1992. Reevaluation of the caudal skeleton
of certain actinopterygian ?shes. III. Salmonidae. Homologization of
caudal skeletal structures. J. Morphol. 214: 187?249.
Bemis, W. E., et al. 2004. Methods for preparing dry, partially articulated
skeletons of osteichthyans, with notes on making ridewood dissections.
Copeia 2004: 603?609.
Berry, F. H. 1959. Young jack crevalles (Caranx species) o? the
southeastern Atlantic Coast of the United States. Fish. Bull. 152:
417?535.
Berry, F. H. 1960. Scale and scute development of the carangid ?sh Caranx
crysos (Mitchill). Quart. J. Florida Acad. Sci. 23: 59?66.
Berry, F. H. 1969. Elagatis bipinnulata (Pisces: Carangidae): morphology of
the ?ns and other characters. Copeia 454?463.
Breder, C. M. 1926. The locomotion of ?shes. Zoologica 4: 159?297.
Britz, R., and Johnson, G. D. 2003. Pardox lost: skeletal ontogeny of Ind-
ostomus paradoxus and its significance for the phylogenetic relationships
of Indostomidae (Teleostei, Gasterosteiformes). Am. Mus. Nov. 3383:
1?43.
Britz, R., and Johnson, G. D. 2005. Leis? conundrum: homology of the
clavus of the ocean sun?shes. 1. Ontogeny of the median ?ns and axial
skeleton of Monotrete leiurus (Teleostei, Tetraodontiformes, Tetraodont-
idae). J. Morphol. 266: 1?10.
Collette, B. B., and Chao, L. 1975. Systematics and morphology of the
bonitos (Sarda) and their relatives (Scombridae, Sardini). Fish. Bull. 73:
516?625.
Dingerkus, G., and Uhler, L. D. 1977. Enzyme clearing of alcian blue
stained whole small vertebrates for demonstration of cartilage. J. Stain
Technol. 52: 229?232.
Donoghue, P. C. J. 2002. Evolution of development of the vertebrate der-
mal and oral skeletons: unraveling concepts, regulatory theories, and
homologies. Paleobiology 28: 474?507.
Fujita, K. 1990. The Caudal Skeleton of Teleostean Fishes. Tokai University
Press, Tokyo.
Gushiken, S. 1988. Phylogenetic relationships of the perciform genera of the
family Carangidae. Jap. J. Ichthyol. 34: 443?461.
Gray, J. 1968. Animal Locomotion. Weidenfeld and Nicolson, London.
Grubich, J. 2003. Morphological convergence of pharyngeal jaw structure
in durophagous perciform ?sh. Biol. J. Lin. Soc. 80: 147?165.
Hilton, E. J. 2002. Osteology of the extant North American ?shes of the
genus Hiodon Lesueur 1818 (Teleostei: Osteoglossomorpha: Hiodonti-
formes). Fieldiana (Zool.), n. ser. 100: 1?142.
Hollister, G. 1941. Caudal skeleton of Bermuda shallow water ?shes. V.
Order Percomorphi: Carangidae. Zoologica 26: 31?45.
Huxley, T. H. 1859. Observations on the development of some parts of the
skeleton of ?shes. Quart. J. Microscop. Sci. 7: 33?46.
Johnson, G. D. 1984. Percoidei: development and relationships. In H. G.
Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr.,
and S. L. Richardson (eds.). Ontogeny and Systematics of Fishes. ASIH/
Allen Press, Lawrence, pp. 464?498.
Johnson, G. D. 1993. Percomorph phylogeny: progress and problems. Bull.
Mar. Sci. 52: 3?28.
Johnson, G. D., and Britz, R. 2005. Leis? conundrum: homology of the
clavus of the ocean sun?shes. 2. Ontogeny of the median ?ns and axial
skeleton of Ranzania laevis (Teleostei, Tetraodontiformes, Molidae). J.
Morphol. 266: 11?21.
Laroche, W. A., Smith-Vaniz, W. F., and Richardson, S. L. 1984. Carang-
idae: development. In H. G. Moser, W. J. Richards, D. M. Cohen, M. P.
Fahay, A. W. Kendall Jr., and S. L. Richardson (eds.). Ontogeny and
Systematics of Fishes. ASIH/Allen Press, Lawrence, pp. 510?522.
Leis, J. M., Olney, J. E., and Okiyama, M. 1997. Introduction to the pro-
ceedings of the symposium ?sh larvae and systematics: ontogeny and
relationships. Bull. Mar. Sci. 60: 1?5.
Liu, C. H. 2001. Early osteological development of the yellow tail Seriola
dumerili (Pisces: Carangidae). Zool. Stud. 40: 289?298.
McKenney, T.W, Alexander, E. C., and Voss, G. L. 1958. Early develop-
ment and larval distribution of the carangid ?sh, Caranx crysos (Mit-
chill). Bull. Mar. Sci. Gulf Caribbean 8: 167?200.
Miller, J. M., and Sumida B, . Y. 1974. Development of eggs and larvae of
Caranx mate (Carangidae). Fish. Bull. 72: 497?514.
Monod, T. 1968. Le complex urophore des poissons te?le?oste?ens. Me?m. Inst.
Fond. Afr Noire 81: 1?705.
Moser, H. G., Richards, W. J., Cohen, D. M., Fahay, M. P., Kendall, A.
W. Jr., and Richardson, S. L. (eds.). 1984. Ontogeny and Systematics of
Fishes. ASIH/Allen Press, Lawrence.
Nelson, G. J. 1969. Infraorbital bones and their bearing on the phylogeny
and geography of osteoglossomorph ?shes. Amer. Mus. Nov. 2394: 1?37.
Patterson, C. 1977. Cartilage bones, dermal bones, and membrane bones, or
the exoskeleton versus the endoskeleton. In S. M. Andrews, R. S. Miles,
and A. D. Walker (eds.). Problems in Vertebrate Evolution. Academic
Press, Lawrence, pp. 77?121.
de Pinna, M. C. C. 1991. Concepts and tests of homology in the cladistic
paradigm. Cladistics 7: 367?394.
Pottho?, T. 1975. Development and structure of the caudal complex, the
vertebral column, and the pterygiophores in the black?n tuna (Thunnus
atlanticus, Pisces, Scombridae). Bull. Mar. Sci. 25: 205?231.
Pottho?, T. 1980. Development and structure of ?ns and ?n supports in
dolphin ?shes Coryphaena hippurus and Coryphaena equiselis (Cory-
phaenidae). Fish. Bull. 78: 277?312.
Pottho?, T., Kelley, S., and Javech, J. C. 1986. Cartilage and bone devel-
opment in scombroid ?shes. Fish. Bull. 84: 647?678.
Reed, D. L., Carpenter, K. E., and deGravelle, M. J. 2002. Molecular
systematics of the jacks (Perciformes: Carangidae) based on mitochon-
drial cytochrome b sequences using parsimony, likelihood, and Bayesian
approaches. Mol. Phylo. Evol. 23: 513?524.
Rieppel, O., and Kearney, M. 2002. Similarity. Biol. J. Linn. Soc. 75: 59?82.
Rosenblatt, R. H., and Bell, M. A. 1976. Osteology and relationships of the
rooster?sh, Nematistius pectoralis Gill. Nat. Hist. Mus. L. A. County
Contrib. Sci. 279: 1?23.
Q3
Q4
188 EVOLUTION & DEVELOPMENT Vol. 9, No. 2, March^April 2007
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
UN
CO
RR
EC
TE
D P
RO
OF
Schnakenbeck, W. 1931. Carangidae. Rep. Danish Oceanograph.
Exped. 1908-10 Mediterranean and Adjacent Seas. Vol. II. Biol. A. 14:
1?20.
Smith-Vaniz, W. F. 1984. Carangidae: Relationships. In H. G. Moser, W. J.
Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L.
Richardson (eds.). Ontogeny and Systematics of Fishes. ASIH/Allen
Press, Lawrence, pp. 522?530.
Smith-Vaniz, W. F., and Staiger, J. C. 1973. Comparative revision of
Scomberoides, Oligoplites, Parona, and Hypacanthus with comments on
the phylogenetic position of Campogramma (Pisces: Carangidae). Proc.
Cal. Acad. Sci. 39: 185?256.
Springer, V. G., and Johnson, G. D. 2000. Use and advantage of ethanol
solution of alizarin red S dye for staining bone in ?shes. Copeia 2000:
300?301.
Starks, E. C. 1911. The osteology and relationships of the ?shes belonging to
the familyCarangidae.Leland Stanford Jr. Univ. Publ. Univ. Ser. 5: 27?49.
Suda, Y. 1996. Osteology and muscular attachments of the Japanese jack
mackerel. Trachurus japonicus. Bull. Mar. Sci. 58: 438?493.
Suzuki, K. 1962. Anatomical and taxonomical studies on the carangid ?shes
of Japan. Rep. Fac. Fish., Prefect. Univ. Mie 4: 43?232.
Tyler, J. C., Johnson, G. D., Nakamura, I., and Collette, B. B. 1989.
Morphology of Luvarus imperialis (Luvaridae), with a phylogenetic
analysis of the Acanthuroidei (Pisces). Smithsonian Contrib. Zool. 485:
1?78.
Whitehouse, R. H. 1910. The caudal ?n of the Teleostomi. Proc. Zool. Soc.
London 1910: 590?627.
Witzell, W. N. 1977. Apolectus niger (family Apolectidae): synonymy and
systematics. Matsya 3: 72?82.
Developmental osteology of carangid ?shes 189Hilton and Johnson
EDE 148(BW
US
 E
DE
 14
8.P
DF
 11
-F
eb
-07
 21
:43
 25
35
90
6 B
yte
s 1
2 P
AG
ES
 n 
op
era
tor
=S
ure
sh
 B
ab
u)
Author Query Form
_______________________________________________________
_______________________________________________________
Dear Author,
During the copy-editing of your paper, the following queries arose. Please respond to these by marking up your proofs with the necessary 
changes/additions. Please write your answers clearly on the query sheet if there is insufficient space on the page proofs. If returning the 
proof by fax do not write too close to the paper's edge. Please remember that illegible mark-ups may delay publication.
 
Journal?? ? ? ? ? EDE
Article??? ? ? ? 148 
Query No. Description Author Response 
Q1 AQ: Please provide manufacturer information for Zeiss Stemi SV 11 dissecting microscope: company name, town, state (if USA) and country. ?
Q2 AQ: Please provide manufacturer information for Nikon COOLPIX 8700: company name, town, state (if USA) and country. ?
Q3 AQ: Please provide volume for reference Berry (1969). ?
Q4 AQ: Please cite Johnson and Britz (2005) in the text or delete from the list. ?
?
? 
?
?
? 
?
?
? 
?
?
? 
?
?
? 
?
?
? 
?
?
? 
?
?
? 
?