Systematics of the Genera Hemicaranx and Atule (Pisces: Carangidae), with an analysis of the classification of the family


Material Information

Systematics of the Genera Hemicaranx and Atule (Pisces: Carangidae), with an analysis of the classification of the family
Physical Description:
xvi, 254 leaves. : illus. ; 28 cm.
Seaman, William, 1945-
Publication Date:


Subjects / Keywords:
Caranigidae   ( lcsh )
Zoology thesis Ph. D
Dissertations, Academic -- Zoology -- UF
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis--University of Florida.
Bibliography: leaves 241-253.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000582624
oclc - 14147593
notis - ADB1001
sobekcm - AA00004935_00001
System ID:

Full Text

Systematics of the Genera Hemicaranx and Atule
(Pisces: Carangidae), with an Analysis of the Classification
of the Family






Dr. Carter R. Gilbert, chairman of my committee, and Mr. Frederick

H. Berry, ichthyologist, deserve special mention for their efforts in

overseeing completion of this dissertation. Above and beyond the

guidance one expects of a chairman, Dr. Gilbert maintained a continual

interest in the work and made completely available his extensive

knowledge of the field and his personal library. His support of

trips to a number of collections -- and his fellowship enjoyed thereon

-- significantly contributed to my graduate program. Fred Berry is a

rare person, dedicated to his profession and the apprentices who come

along; he initially suggested the work on Hemicaranx, and generously

provided personal notes and data that would assist this work. His

extensive knowledge of the Carangidae has provided a valuable review


The manuscript has benefited from review by my committee, Drs.

Brodkorb, Nicol, and Carr, for whose efforts I am grateful.

The following colleagues provided material assistance in obtaining

specimens on loan or allowed me to use their institutional facilities

(see abbreviations section): Dr. James Atz, AMNH; Drs. James Bohlke

and James Tyler, ANSP; Dr. William Eschmeyer, CAS; Mr. Loren Woods,

FMNH; Mr. Robert Topp, FSBC; Dr. Ralph Yerger, FSU; Mr. Charles Dawson,

GCRL; Drs. Robert Lavenberg and Camm Swift, LACM; Drs. Carl Hubbs and

Richard Rosenblatt, S10; Mr. George Miller, Dr. Robert V. Miller and

Mr Frederick H. Berry, TABL (now Southeast Fisheries Center); Dr. Royal


Suttkus, TU; Dr. Boyd Walker, UCLA; and Drs. Ernest Lachner and

Victor Springer, USNM, (now National Museum of Natural History).

Also, Dr. D. E. McAllister, NMC, loaned specimens. Mme. M. Bauchot,

MNHN, and Dr. M Boeseman, RMNH, generously allowed the use of facilities

at their respective institutions. Dr. Boeseman further permitted a loan

of type material. Dr. Jdrgen Nielsen, Universitetets Zoologische Museum,

Copenhagen, provided data on the type of Atule djedaba. To Dr. Peter J.

P. Whitehead, BMNH, I extend similar thanks for the loan of specimens,

use of facilities, and provision of data; furthermore, his friendship

since our meeting has been one of the treasured experiences of prepar-

ing this dissertation.

Dr. John Randall, Bernice Bishop Museum, Hawaii, and Mrs. Margaret

Smith, J. L. B. Smith Institute of Ichthyology, South Africa, provided

information on their collections.

Technical assistance was provided by University of Florida artists

Mr. Paul Laessle and Ms. Margaret Estey who put aside busy schedules

to offer suggestions about the drawings. Mr. Russell Parks, Image

Designs, Gainesville, took all the photographs.

To my friends -- among whom I am fortunate enough to include my

parents and my loving wife,Carol -- who helped, a quiet word of thanks.





LIST OF TABLES ... ... . . viii


ABBREVIATIONS. ...... . . .xii

ABSTRACT . . ... .. .. xv


METHODS .. . . 5

Numerical Taxonomy. ..... . .. 7
Preparation of Skeletal Material. . 8


A Review of the Suzuki Classification of
Japanese Carangidae . . ... 12
Primitive characters of the Carangidae . .. 13
Numerical Taxonomy of the Japanese Carangidae .. 16


Infraspecific variation . . 21
Descriptive Osteology of Hemicaranx . ... 22



Hemicaranx Bleeker .. ............ 50
Hemicaranx amblyrhynchus (Cuvier) ............. 63
Hemicaranx bicolor (Gunther). . .. 93
Hemicaranx zelotes Gilbert . .... 99
Hemicaranx leucurus (GUnther) . .123


Atule Jordan and Jordan .. . .. 130
Atule mate (Cuvier) . . 150
Atule kalla (Cuvier)... .. ....... 156
Atule macrurus (Bleeker)...... 160
Atule djedaba (Forskal) . .... 163
Atule malam (Bleeker) . . 167

APPENDICES . ... ... 170

Appendix 1. Tables ...... . 171
Appendix II. Osteological Figures . .. 208
Appendix III. Skeletal Material Examined . 235

GLOSSARY . . ... ......... .238


BIOGRAPHICAL SKETCH. . ... ..... 254


Figure Page

1. Phylogeny of Japanese Carangidae . ....... 14

2. Phenetic dendrogram of Indo-Pacific Carangidae .. 17

3. Phenetic dendrogram of genera of Carangidae sharing
a matching coefficient of at least .75 with
Hemicaranx . . 61

4. Distribution of Hemicaranx in the Atlantic and Pacific
Oceans . . . 79

5. Length of upper caudal fin lobe in populations of
Hemicaranx amblyrhynchus from the Western Atlantic
Ocean and H. bicolor from the Eastern Atlantic .. 82

6. Variation in number of fin rays in the soft dorsal
fin of Hemicaranx . .. .. 83

7. Variation in number of fin rays in the soft anal
fin of Hemicaranx .. ... . .. 84

8. Adult Hemicaranx from the Atlantic Ocean ... .87

9. Interspecific variation in number of scales in
curved lateral line of Hemicaranx ... . 88

10. Winter Atlantic Ocean equatorial surface currents 91

11. Interspecific variation in ratio of straight to
curved lateral line lengths in large adult Hemicaranx 107

12. Number of teeth in the premaxillary bone of
Hemicaranx from Panama Bay, Panama .. . 108

13. Number of teeth in the dentary bone of Hemicaranx
from Panama Bay, Panama ..... . .... 109

14. Interspecific variation in number of teeth in large adult
specimens of Hemicaranx . .. 110

15. Length of the pectoral fin in Hemicaranx from Panama
Bay, Panama . . ... .... .Ill

16. Length of the pelvic fin in Hemicaranx from Panama
Bay, Panama . . ... .... .112

17. Lateral outlines of anterior caudal peduncle scutes
in Eastern Pacific Hemicaranx . .. 114

18. Width of the anterior peduncle scute in Hemicaranx
from Panama Bay, Panama . .... 115

19. Juvenile specimens of Hemicaranx . .... 117

20. Adult Hemicaranx from the Eastern Pacific Ocean .... 119

21. Lateral outlines of anterior ends of articulated
premaxillary and dentary bones in Eastern Pacific
Hemicaranx . . ... ... 121

22. Interspecific variation in width of the anterior
peduncle scute in large adult Hemicaranx ... 122

23. Phenetic dendrogram of genera of Carangidae
sharing a matching coefficient of at least
.75 with Atule . ........ .... 134

24. Ratios of straight to curved lateral line
lengths in the species of Atule . .. 139

25. Depth of head in Atule . .... 140

26. Species of Atule . .... 142

27. Width of the anterior caudal peduncle scute
in the species of Atule . .... 143

28. Number of premaxillary teeth in the species of
Atule characterized by uniseriate dentition ...... 144

29. Relative lengths of ultimate (terminal) and pen-
ultimate soft dorsal fin-rays in Atule ... 145

30. Interspecific variation in number of fin rays in
the soft dorsal fin of Atule . ... 146

31. Interspecific variation in number of fin rays in
the soft anal fin of Atule . .. 146

32. Interspecific variation in number of scales in
curved lateral line of Atule . .. 147

33. Interspecific variation in number of scutes in
straight lateral line of Atule . .. 147

34. Interspecific variation in width of the anterior
peduncle scute in Atule . .... .148

35. Distribution of the species of Atule .. 149




Geographic distribution of the nominal genera of Carangidae


. 172

2. Characters coded in numerical taxonomy of
Carangidae . . ... ....... 173

3. Character state-operational taxonomic unit matrix
for Carangidae .. . . 174

4. Coefficients of association for carangid operational
taxonomic units from Japan . .... 176

5. Second generation matrix of association coefficients,
calculated for cluster stems and individual OTU's
incorporated in Japanese carangid cluster analysis ...... 177

6. Distribution of primitive skeletal character states
in Japanese Carangidae, plus Hemicaranx. . .. 178

7. Morphometric values for Hemicaranx amblyrhynchus
from Western Atlantic localities . .. 179

8. Meristic values for Hemicaranx amblyrhynchus from
Western Atlantic localities . .... ... 182

9. Morphometric values for Hemicaranx bicolor from
three West African localities . .... 183

10. Meristic values for Hemicaranx bicolor from three
West African localities . ..... ... 185

11. Comparative morphometric values for two species of
Hemicaranx from Panama Bay, Panama. . .... 186

12. Comparative meristic values for two species of
Hemicaranx from Panama Bay, Panama . .... 189

13. Morphometric values for Hemicaranx zelotes from
the lower Pacific coast of Baja California, Mexico ...... 190

14. Meristic values for Hemicaranx zelotes from the
lower Pacific coast of Baja California . .... 191

15. Number of dorsal fin rays in four species of
Hemicaranx . . ... ....... 192


16. Number of anal fin rays in four species of
Hemicaranx . . 193

17. Number of pectoral fin rays in four species of
Hemicaranx . ... 194

18. Number of scales in the curved lateral line of
Hemicaranx . . .. 195

19. Number of scales in the straight lateral line of
Hemicaranx . . 195

20. Number of scales in the curved lateral line of
Atule . . ... .197

21. Number of scutes in the straight lateral line
of Atule . . .197

22. Coefficients of association between Hemicaranx
OTU and OTU's from Japan .. . 199

23. Coefficients of association between Chloroscombrus
OTU and OTU's for which character state information
is available . ... ... 200

24. Atlantic Ocean surface temperatures .. 201

25. Comparison of diagnostic characters of the species
of Hemicaranx . .... ..... 202

26. Morphometric values for the species of Atule 203

27. Meristic values for the species of Atule ... 205

28. Number of teeth in species of Atule with
uniseriate dentition .. .. . 206

29. Number of dorsal fin rays in the species of Atule 206

30. Number of anal fin rays in the species of Atule .... .206

31. Comparison of diagnostic characters of the species
of Atule . . ... ...... 207


Figure Page

1. Neurocranium of Hemicaranx zelotes 210

II. Dorsal views of anterior edge of dorsal surface
of ethmoid bone in four species of
Hemicaranx 212

III. Suborbital series in Hemicaranx zelotes 212

IV. Dorsal view of two representative suborbital
shelves in Hemicaranx amblyrhynchus 212

V. Lower jaw of Hemicaranx zelotes ....... 2 214

VI. Outline of posterior edge of left dentary in
Hemicaranx... 214

VII. Lateral view of upper jaw of Hemicaranx zelotes 216

VIII. Outline of dorsal edge of left premaxillary in
Hemicaranx. . . .216

IX. Lateral view of hyomandibular bones of Hemicaranx
zelotes . . .. .218

X. Lateral outline of symplectic bone in Hemicaranx 218

XI. Lateral outline of pterygoid bone in Hemicaranx 218

XII. Lateral view of opercle, subopercle, and inter-
opercle in Hemicaranx zelotes . .. 220

XIII. Adpharyngeal view of basihyal bone and branchial
elements of Hemicaranx zelotes .. .. .222

XIV. Lateral views of hyal bones of Hemicaranx zelotes 224

XV. Outlines of selected hyal elements in Hemicaranx 26

XVI. Appendicular skeleton of Hemicaranx zelotes 228

XVII. Outline lateral view of postcranial axial and
medial skeleton, exclusive of caudal skeleton,
in Hemicaranx zelotes .. . .230

XVIII. Pterygiophore lepidotrich articulation in
Hemicaranx zelotes . .... .232

XIX. Lateral outlines of the distal ends of anal
pterygiophores in Hemicaranx . .... .232

XX. Representative vertebrae in Hemicaranx zelotes 232

XXI. Lateral view of caudal skeleton of Hemicaranx
zelotes . . .. 234

XXII. Ontogenetic fusion and development in Hemicaranx
zelotes and Chloroscombrus chrysurus. ...... 234



AMNH American Museum of Natural History, New York City
ANSP Academy of Natural Sciences of Philadelphia
BMNH British Museum (Natural History), London
CAS California Academy of Sciences, San Francisco
CAS-SU Stanford University collection, now at CAS
CM Charleston Museum, Charleston, South Carolina
FMNH Field Museum of Natural History, Chicago
FSBC State of Florida Marine Research Laboratory,
St. Petersburg
FSU Florida State University, Tallahassee
GCRL Gulf Coast Research Laboratory, Ocean Springs,
LACM Los Angeles County Museum of Natural History
MNHN Museum National d'Histoire Naturelle, Paris
NMC National Museum of Canada, Ottawa
RMNH Rijksmuseum van Natuurlijke Historie, Leiden
S1I Scripps Institution of Oceanography, La Jolla
TABL Tropical Atlantic Biological Laboratory, Miami
TU Tulane University, New Orleans
UBC Institute of Fisheries, University of British Columbia,
UCLA Department of Zoology, University of California, Los
UF Florida State Museum, University of Florida, Gainesville
USNM United States National Museum, Washington, D. C.

Bones, By Region


BOC basioccipital
BS basisphenoid
DSO dermosphenotic
E ethmoid
EOC exoccipital
EPO epiotic
F frontal
L lacrymal
LE lateral ethmoid
N nasal
OPS opisthotic

P parietal
PRO prootic
PS parasphenoid
PTO autopterotic
PTS pterosphenoid
PV prevomer
S sclerotic
SPH autosphenotic
SUB suborbital
SUO supraoccipital


Oromandibular region

Upper jaw bones

MX maxillary
PMX premaxillary
SMX supramaxillary

Lower jaw bones

AN angular
AR articular
D dentary

Hyoid region

B branchiostegal rays
BH basihyal
CH ceratohyal
EH epihyal
HM hyomandibular
IH interhyal
lOP interopercle
LHH lower hypohyal
MSPT mesopterygoid
MTPT metapterygoid
OP opercle
PAL palatine
POP preopercle
PT pterygoid
Q quadrate
SOP subopercle
SY symplectic
UH urohyal
UHH upper hypohyal

Branchial region

BB basibranchial
CB ceratobranchial


EB epibranchial
HB hyopobranchial
PB pharyngobranchial

Appendicular Skeleton

Pectoral girdle

CL cleithrum
CO coracoid
LEP lepidotrichs
PCL postcleithrum
PTM posttemporal
R radials
SC scapula
SCL supracleithrum

Pelvic girdle

BPT basipterygium
LEP lepidotrichs

Axial Skeleton

CV caudal vertebrae
EPR epipleural rib
PCV trunk (precaudal) vertebrae
PR pleural rib

Medial Skeleton

DPT distal pterygiophores
LEP lepidotrichs
PD predorsals (supraneurals)
PPT proximal pterygiophores

Caudal Skeleton

CAP antepenultimate caudal vertebra
CP penultimate caudal vertebra
EP epural
HS hemal spine
HY hypural
NS neural spine
PCR principal caudal ray
SCR secondary caudal ray
UN uroneural
UR urostylar (terminal) vertebra


Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy

Systematics of the Genera Hemicaranx and Atule
(Pisces: Carangidae), with an Analysis of the Classification
of the Family


William Seaman, Jr.

August, 1972

Chairman: Carter R. Gilbert
Major Department: Zoology

The carangid fish genus Hemicaranx Bleeker is an uncommon compon-

ent of the subtropical and tropical ichthyofauna of the Atlantic and

Eastern Pacific Oceans. Analysis of specimens from throughout the

range for external morphological and osteological characters was per-

formed to resolve the systematic status of the nominal species of

Hemicaranx. Based on examination of morphometric, meristic, and

osteological characters, four species are recognized: H. amblyrhynchus

(Cuvier), a wide-ranging Western Atlantic shore species that is document-

ed to be estuarine dependent; H. bicolor(Ginther) of the Eastern

Atlantic, closely related to amblyrhynchus but differing with regard

to number of dorsal and anal rays, length of caudal lobes, and several

osteological characters; and H. zelotes Gilbert and H. leucurus (Gunther),

which occur sympatrically over much of their ranges along the Eastern

Pacific coast from Mexico to Peru. The latter two species are distin-

guished on the basis of tooth morphology, body bars, caudal scute width,

and, in larger individuals, pectoral fin length. H. zelotes and H.

leucurus are more closely related to each other than either is to H.

amblyrhynchus and H. bicolor.

To describe the osteology of Hemicaranx, two existing methods for

obtaining skeletal material from preserved specimens were combined for

the first time. Thus, larger preserved material was heated in enzyme-

based detergent solution and rinsed with ammonia, thereby speeding up


Because the species of Hemicaranx and Atule are sometimes combined

in a single genus, and since a number of workers have questioned their

relationship, the hypothesis that the two genera are distinct but close-

ly related was explored. Osteological data from the present study were

combined with a synthesis of the osteological catalog of Suzuki to assess

the relationship of Hemicaranx to the Indo-Pacific Carangidae. Based

on the numerical taxonomic analysis conducted, I conclude that Hemicaranx

and Atule are generically distinct, though very closely related. The

phenetic classification of this study confirms to a large degree the

phylogeny proposed by Suzuki for Japanese carangids.

Study of limited geographic and ontogenetic samples of Atule

provides the basis for review of the genus. Five valid species, all

in the Indo-Pacific basin, are recognized: A. kalla, A. mate, A. malam,

A. djedaba, and A. macrurus. Diagnostic characters include dentition,

lateral-line ratios, and number of lateral-line scutes.

Ecologically, both Hemicaranx and Atule are characterized by

early juvenile stages that are commensal with jellyfishes. The possibil-

ity of transoceanic dispersal by currents is seemingly confirmed by the

transport of H. bicolor from Africa to northeastern South America.


The teleostean fish family Carangidae, whose members are commonly

known as the jacks, pompanos, and scads, is well known because of the

commercial and/or sport fisheries supported by a relatively small number

of its species. Many members of this distinctive, primarily circum-

tropical, marine family are still poorly known systematically, however,

and as Berry has noted (1968: 164) the morphological characteristics

and limits of a number of carangid genera are inadequately defined,

pending thorough analysis of the species involved. Such is the case

for Hemicaranx Bleeker, a small, uncommonly collected genus whose

species are found in the tropical and subtropical coastal waters of

the Atlantic and eastern Pacific Oceans. With the accumulation of pre-

served specimens of the nominal species in institutional collections,

it is now possible to revise the species of Hemicaranx. Based on

examination of (1) intraspecific variation of external morphology,

including several characteristics that have not been included in

earlier accounts of the species, and (2) osteology, which heretofore

has not been studied in Hemicaranx, the limits of the genus are also

defined in this study.

In seeking to identify the carangid genera with which Hemicaranx

might share a more or less common ancestor, it became apparent that

the supposedly phylogenetic classification of the Carangidae rests

mainly on subjective grounds. Without an abundance of fossil evidence

to provide an objective basis for definition of "primitive" characters,


students of this and many other groups have tended to base their class-

ifications on either intuition or the assumption that the most widely

shared characters are most typical of ancestral forms. Indeed, taxon-

omically significant characters have frequently (but uncritically)

been assumed to be of significance in defining phylogenetic trends; the

danger of such a practice was pointed out by Gilbert and Bailey (1972:

9). Sokal and Sneath (1963: 67) extensively discussed the problems at-

tendant to inference of phylogeny from affinity (morphological resemb-

lance, etc.) of organisms. To identify the genera of closer affinity

to Hemicaranx two alternative approaches may be employed:

1. Overall comparison of genera, as one might do in "keying-out"

a specimen, utilizing perhaps only a few characters that are more

"significant" taxonomicc or phylogenetic?) in constructing a phy-


2. Comparisons based on large numbers of non-weighted, randomly

selected characters that result in an objective description of

phenetic resemblance.

The former approach is typical of the literature dealing with the

Carangidae; of recent import is the classification of the traditionally

recognized and accepted generic taxa of Japanese carangids by Suzuki

(1962), who based a phylogeny upon osteological characters. Also, on

the basis of "weighted" characters reflecting general morphological

similarities, the genus Atule of the Indo-Pacific basin was referred

to as a possible "close relative" of Hemicaranx by Nichols (1942b: 229).

Indirect comment on the affinity of Hemicaranx to Atule was provided

by Fowler, who included species of both genera in the invalid genus

Alepes. Preliminary examination and comparison of nominal species of

Atule with Hemicaranx confirms these appraisals.

Because Atule is poorly known and its species frequently confused,

and because of relatively close affinity to Hemicaranx, its species --

known from limited collections -- are reviewed. The relationship of

Atule to Hemicaranx will be assessed as part of the overall review of

generic classification presented in this study.

The second approach is based on the principles of numerical taxonomy.

Because these techniques have not previously been applied to the Carangid-

ae, or many other teleosts, the procedures of Sokal and Sneath (1963), on

which they are based, are briefly summarized and are then employed to

identify those genera to which Hemicaranx is related. In addition to

providing an alternate method of identifying relative affinities of gen-

era, numerical taxonomy will be used in this study to generate a phenetic

classification of the Indo-Pacific Carangidae. As a means of evaluating

the phylogeny proposed by Suzuki (1962), the osteological characters he

used will be incorporated in this alternate classification insofar as

possible. In reviewing Suzuki's work it was necessary to reconcile hypo-

thesized phyletic trends with the fossil record; a summary of primitive

carangid characters is essential to this task.

As with many marine teleost families, the Carangidae are more

diverse in the Indo-Pacific region. Because more genera are present

there than in the Atlantic and Eastern Pacific Oceans, it is likely

that the Indo-Pacific basin is the center of origin of the Carangidae

(Table 1). Of interest is the apparent failure of several genera from

the Indo-Pacific to enter the Atlantic or cross the Eastern Pacific past

Hawaii, perhaps partly because of either relative age of genera or their

relative rates of dispersal (see Table 1). Also of note is that several

genera (including Hemicaranx) are found only in the Atlantic and

Eastern Pacific basins. This may be due to a) lack of critical study

and definition of genera (Mansueti, 1963: 55; Berry, 1968: 164),

resulting in simple unresolved synonymies, or b) evolution of new

taxa (genera) from some genera that did indeed reach the Atlantic and/or

Eastern Pacific either through the Tethys Sea of pre-Miocene times, or

around the southern tip of Africa, or across the Pacific Ocean. The

numerical taxonomy generated in this study may also prove useful in

assessing the various evolutionary histories of the genera of Carangidae.

In particular, it will be employed to comment on the affinities of



Counts and measurements of external morphological characters of

fish specimens examined were based as much as possible on standard

ichthyological procedure. For the most part, meristic and morphometric

data are expressed in terms of the definition of Hubbs and Lagler (1958:

19-26). However, the unique characteristics of certain Carangidae

necessitate the use of additional or modified counts and measurements;

in this regard, I have attempted to use those characters already defined

in the literature dealing specifically with carangid fishes (i.e.,

Berry, 1959; Williams, 1959; and Berry, 1968). Definitions of terms

from the literature, as well as terminology unique to this study, are

presented in the Glossary. All measurements were taken using dial

calipers; distances greater than 100 mm were read to the nearest


Deduction of evolutionary history usually is based on the assump-

tion that degree of relationship is positively correlated with degree

of resemblance. To strengthen the conclusions based on such comparisons

the number of characters examined may be increased to reduce misleading

interpretation of convergence of certain characters. Although external

morphology is most commonly employed, osteological features may be also

used to expand the evidence of taxonomy, and to construct phylogenies

based on degree of resemblance, Indeed, the supposed "conservative"

nature of osteological characters has prompted workers to accept them

as preferred kinds of evidence. (Norden [1961: 683], for example,


acknowledged this, but he also employed developmental stages and.other

morphological aspects.) However, the occurrence of significant infra-

specific variation of certain osteological characters in some groups of

fishes (Schleuter and Thomerson, 1971) demonstrates the fallacy of a

typological approach to osteology. The need to account for infra-

specific osteological variation before incorporating such information

in taxonomic definitions is apparent.

Analysis of infraspecific osteological variation was based on exam-

ination of up to five specimens of closely similar standard length (65-

70 mm). Ontogenetic change was studied in specimens ranging from 33 to

150 mm SL. Osteological observations were made primarily on cleared and

stained specimens, but other preparations were also employed, as dis-

cussed below. Skeletal material examined is listed in Appendix III.

As nearly as possible, osteological terminology conforms to that

used in more recent publications on the Carangidae (i.e., Suzuki, 1962,

Berry, 1969). However, all usage has been reconciled with accepted

ichthyological literature; pertinent in this regard are Harrington, 1955

(osteocranium), Smith and Bailey, 1961 and 1962 (dorsal skeleton, sub-

ocular series), and Gosline, 1961 (caudal skeleton). Terminology at

variance with Suzuki (1962) is so noted in the text. A listing of

skeletal elements is provided in the Abbreviations section. In the

osteological section, a detailed account of the location, orientation,

and morphology of each skeletal element is provided for one species of

Hemicaranx, namely H. zelotes. Following the description of each bone,

any interspecific differences are noted.

Material examined in the course of this study is listed in each

species account. For each lot of specimens studied, data are presented

in the following format: Institutional catalog number; number of speci-

mens and size range (mm SL), in parentheses; and locality data (country,

state or province, county if in U. S., geographic locality, latitude, longi-

tude, depth, cruise, collector, date). Institutional abbreviations are

listed at the front of the text.

Outline drawings of material were made with a Wild camera lucida,

modelled in pencil, and ink drawings were then executed.

Statistical treatment of data was carried out by means of a Monroe

Epic 3000 Calculator in the computation of descriptive statistics, and

a Monroe Epic 1665 Calculator in the generation of t-statistics for

unequal sample size and unequal standard deviation (Snedecor, 1956: 97-

98). Comparison of larger numbers of samples was effected using the

graphical approximation to a multiple-comparison test of Eberhardt (1968).

The advantages of this technique, especially preservation of the stated

level of significance, are discussed by Eberhardt; significance levels

stated in figure legends of this text are derived from Eberhardt (1968:

Figure 2). Inspection of data reveals that many characters follow allo-

metric growth curves; for such characters restricted straight-line por-

tions of the curve were compared for different samples of similar-sized

individuals. Ontogenetic growth is discussed in the species accounts.

Numerical Taxonomy

As employed by McAllister (1966) the techniques of Sokal and Sneath

(1963) have been shown to be of utility in assessing and establishing

a phenetic classification of fishes. McAllister (1966: 227-229) pro-

vides a summary of the methods of coding characters and calculating and

tabulating simple matching coefficients; however he does not extend the

techniques of Sokal and Sneath beyond a coefficient matrix, nor does he

subject his data to standard cluster analysis to plot a dendrogram.

Therefore, a summary of the techniques of Sokal and Sneath herein em-

ployed is provided:

1. Selection of characters.

2. Coding. Each character was recorded as being either present

or absent for each operational taxonomic unit (OTU)(i.e., generic taxon).

A list of the character states is provided in Table 2; a matrix of char-

acter states appears in Table 3.

3. Calculation of matching coefficients of association, based on

the formula Ssm= m/n, where Ssm is the coefficient, m is the number of

character states shared between OTU's, and n is the total number of char-

acter states compared. Matching coefficients are listed in Table 4.

4. Cluster analysis. As evaluated by Sokal and Sneath (1963: 189)

the "weighted pair group method" of cluster analysis, using averages to

calculate new similarity coefficients for each generation, results in the

least distortion of dendrograms. This technique is herein employed to

plot dendrograms; new members are weighted as equal to the sum total of

old cluster group members (Sokal and Sneath, 1963: 190-191). A second

generation matrix is illustrated in Table 5.

Preparation of Skeletal Material

Investigations of teleostean osteology have usually been based on (1)

X-rays of specimens, (2) cleared and stained material, and/or (3) dry

bones that have been cleaned of soft tissues. Limitations of each tech-

nique exist: X-rays obscure 3-dimensional detail and prevent direct

handling of bones; very large specimens frequently fail to clear even

after months of enzyme digestion, and certain groups are resistant to
the process (Miller and Van Landingham, 1969: 829); dissection and re-

moval of bones from preserved specimens may be prohibitively time

consuming, whereas fresh material from which dry bone preparations may

be obtained by maceration is frequently not readily available. Thus,

the osteological characters of many fish groups have not been incorporated

in systematic studies.

Collections of comparative skeletal material of the species of

Hemicaranx are extremely limited. Because the enzyme method of Taylor

(1967) for clearing and staining small vertebrates did not yield satis-

factory results on carangid fishes greater than 100 mm SL, I employed

two recently developed techniques for the preparation of dry bones from

preserved material to examine the osteology of individuals of Hemicaranx

above 100 mm SL. I found, however, that Konnerth's (1965) method of pre-

paring ligamentary articulated specimens consumed an inordinate amount of

time in the skinning and removal of tissue from specimens. The possibility

of skeletal damage by chlorine, which must be employed in this method, is

also a drawback (Ossian, 1970: 199). Meanwhile, the technique of Ossian

(1970) for specimen disarticulation using enzyme-based laundry "pre-

soakers" consistently resulted in partial disintegration of superficial

bones, especially dermal elements of the jaws and opercular series, before

complete disarticulation could occur. To avoid these difficulties, I

combined complementary aspects of each technique.

Preserved whole carangid specimens (150-200 mm SL) were transferred

from alcohol storage into "Biz" solution and maintained at 700 C, follow-

ing the procedure of Ossian (1970). Within 24 to 96 hours after initial

immersion, deterioration of superficial skin and membranes takes place,

although,because it precedes bone damage, such deterioration is a useful

signal of imminent osteological disintegration or alteration. At this

point, then, more superficial skeletal elements may be easily removed

before they are altered, either individually or as a unit, and the remain-

der of the skeleton and attached flesh returned to a fresh "Biz" solution.

After brief additional soaking the deeper cranial and axial muscle masses

are easily split off in chunks, thus effecting considerable time-savings

over the method of Konnerth (1965). Exposure of bare bone to "Biz" solu-

tion necessitates termination of soaking, but the remaining soft tissues

may now be readily removed -- especially if the specimen is soaked in

ammonia after rinsing, as discussed by Konnerth (1965: 328). By this

time, too, all preservative has been washed out of the specimen, and

maceration in water may also be employed to remove remaining tissue.

With judicious combination of the techniques of Konnerth (1965)

and Ossian (1970) it is now possible to prepare adequate amounts of un-

damaged osteological material from preserved fish specimens. With dis-

section, units of the skeleton may be retained in articulated condition.


Classifications are frequently based on a relatively low number

of taxonomic characters. For example, ever since Bleeker's (1862: 135-

138) initial taxonomic distinction of several carangid genera on the

basis of dentition, distribution of teeth has been accorded special

taxonomic importance in the classification of Carangidae. Over the years

a small number of additional characters have been incorporated into

classifications as a means of defining carangid taxa: presence or ab-

sence of lateral-line scutes, for example, has been utilized in the

establishment of subfamilies, just as presence or absence of detached

finlets has been ascribed value in generic diagnoses. More extensive

subsequent description of overall external morphology of species initi-

ally distinguished by one or a few diagnostic characters has, for the

most part, confirmed the initial taxonomic interpretations, thus imply-

ing that groups of characters vary in a correlated fashion. This con-

clusion has apparently served as a "carte blanche" for the weighting of

certain characters as more important than others in establishing taxo-

nomies. That is, suites of correlated characters may be incorporated

into taxonomy as a group (i.e., the "single adaptive complex" of Mayr,

et al., 1953: 123) by weighting a single member of the suite and letting

it represent the group, thus eliminating the need to continually repeat

character state information.

Besides the use of weighted characters -- which are not always

documented 1) to be representative of a suite or 2) to be primitive --

speculation about evolutionary trends has usually been based on the

assumption that closely related taxa share relatively more characteristics

than do more distantly related forms. For the Carangidae, for example,

Ginsburg (1952) ascribed phylogenetic relationship on the basis of ex-

ternal morphological similarity. Most notable in this regard is the

phylogeny proposed by Suzuki (1962) for Indo-Pacific carangids based

upon an extensive catalog of osteological features.

A Review of the Suzuki Classification
of Japanese Carangidae

In a comparison of Indo-Pacific carangids from Japan (Table 1),

Suzuki (1962) provided descriptions of variable detail for over 70

osteological features. Out of 34 characters demonstrated to be diagnostic

for all genera, Suzuki (p. 130) listed ten that are "significant for dis-

closing their phylogeny." Based on these characters a progenitor is

hypothesized, and extant forms are compared to the "ideal" in an effort

to delineate the evolutionary trends of subfamilies. In a subsequent

discussion of evolution within subfamilies, Suzuki cited various diagnostic

characters as evidence for derivation of phyletic units. However, in his

discussion Suzuki ,did not document many of the evolutionary trends he

hypothesized; if he had a rationale for a widely opened myodome being

primitive (p. 66), for example, he did not express it.

Also, out of many so-called primitive characters Suzuki chose to

weight some as being of special significance in phylogeny. In his dis-

cussion of the suspensorium and opercular apparatus, for example (p. 87),

he listed eleven characters significant to classification; shape of

pterygoi,d and height of apparatus are stated to be of particular

importance. Again no rationale was given for the differential phylo-

genetic significance.

The "disposition of genera in conformity with their degree of

differentiation" proposed by Suzuki (p. 133) is illustrated in Figure

1. Essentially this phylogenetic tree is based on documented and un-

documented suppositions that, of the scores of features examined,

but a relative few are of utility in deducing phylogeny.

Primitive Characters of the Carangidae

The Carangidae first appear in the fossil record during the Eocene,

and they are thought to have arisen from a dinopterygoid beryciform

close to the genus Aipichthys (Patterson, 1964: 398). During the course

of evolution the Carangidae have lost the following characters that

are still found in Aipichthys: eight branchiostegal rays (reduction to

seven), orbitosphenoid, toothed endopterygoid, fewer dorsal spines and

no free spines in front of the dorsal, 17 branched caudal rays (reduction

to 15) (Patterson, 1964: 397). Characters shared between Aipichthys

and the Carangidae include: high supraoccipital crest; upturned,

protrusible mouth; absence of ornamentation on cranial bones; single,

elongate supramaxillary; number of vertebrae (10 + 15); form of

cleithrum (lengthened and broadened ventrally); form of coracoid

(enlarged); long dorsal and anal fins with a few spines and with

elongate anterior rays; cycloid scales; and especially deeply forked

caudal-ray bases (Patterson, 1964: 397, 469-470).

Gregory (1933: 300-303) cited as primitive characters for the

Carangidae, a low number of vertebrae (24-26), spinous dorsal and anal

-J <
- LJ
(_ LU


F0 -
C/ 0

0A <
Z (I Z -1
o LU U, 0 0 0 < -D (<
O 2
/) Z U 0 -
- <0 3 F- -I <
0 I- 0 =3 < 3
J < 0 I- a -
-1 Z < <
c, I



Z ) LI

...... ........ ........

fins that are neither reduced nor separated, non-broadened opercular

(antero-posteriorly), and less elongate head. He also stated that in

more primitive carangids the supraoccipital-frontal crest is steeper

and higher, the opercle relatively deep and short, the mouth small,

with a long ascending process of a protrusible premaxillary, the

quadrate articular joint moderately far forward, and the body is

deeper ("ovate to orbicular"). Presumably, Gregory based these

trends on examination of Aipichthys, which he referred to as a "deep-

bodied Cretaceous form" that may be the real ancestor of the Carangidae

(1933: 300).

If these substantiated primitive characters are compared with

those used by Suzuki, one can see that he was correct in weighting

certain of the characters he employed as phylogenetic evidence.

Included are the primitive character states of (1) eight branchio-

stegal rays, (2) a supramaxillary, (3) a high supraoccipital crest,

and (4) a protractile premaxillary. In addition, (5) the absence of

scutes, specialized structures appearing only in some members of the

family, and (6) a feebly developed first hemal spine appear to be

typical of primitive carangids and apichthyids. Finally, the (7)

expansion of the post-maxillary process as a bracing supportive struct-

ure, ancillary to the development of a protrusible mouth as an evol-

utionary advancement (Patterson, 1964: 456), appears to be a valid

trend in the group. Data of Suzuki for these few characters, when

tabulated after the manner employed by McAllister (1966: 230-231),

give some indication of the degree of resemblance of extant carangids

to ancestral forms (Table 6). Because data on only six primitive

characters are available for all genera in Suzuki, little significance

is attached to the tabulated summaries. Indeed, all genera have

between -two and four primitive character states present. Clearly,

further appraisal of primitiveness, based on documented literature

accounts of fossils, must await additional study of appropriate

characters, including studies of external morphology, a task acknowled-

ged by several workers.

In addition to a low number of documented primitive characters,

Suzuki also incorporated into his phylogenetic classification weighted

characters for which the primitive state is conjectural. Hypothe-

sized primitive characters that were weighted include (1) rostrum

wide and short, (2) myodome opening wide, (3) ceratohyal window

wider, (4) urohyal shorter, and (5) olfactory cavity absent. Finally,

numerous other characters are unweighted in this study. As a result,

even though Suzuki reviewed scores of osteological features, only

a small number of diagnostic generic characters are incorporated

into his "hybrid-phylogeny," one that is inconsistent in employment

of characters and their weighting.

Numerical Taxonomy of the Japanese Carangidae

The extensive osteological catalog for the Carangidae provided

by Suzuki (1962) may be incorporated into an alternative scheme of

classification, based on the techniques of numerical taxonomy, in

which all characters are weighted equally. (The large number of

characters examined by Suzuki are not equally nor uniformly described

for all species in his discussion; this analysis is based only on the

characters completely cataloged in his paper.) Equal weighting

(Sokal and Sneath, 1963: 118-120) is employed as an alternative to the

dilemma of allocating differential weights to characters that may or

= V)

Z V) )) 00 c )
I-I ol < Z Q
0 0< x Ce ccl <

F I- 0-C0 Z o 0
-1 < < 0 <
< <

0 U) U)
00 w
< 0 n-
= w 0- =
I- < < U
< -I 0c Z LU. LU Q:
C3 U)1 01-

- V ) 1- C
CL. t/ U

< < i -z
X<3 w c <0
LU < LU a: =
S U Z U) 1- t-

Figure 2. Phenetic dendrogram of Indo-Pacific

70 o

may not be classified as to primitivenesss." As discussed above, only

a small number of characters may unquestionably be described as

primitive for the Carangidae; many others may or may not be. (An infin-

ite number of weightings could be assigned, therefore, with evaluation

largely a matter of subjective preference.)

As suggested by McAllister (1966: 227), the minimum of 40 charac-

ters were coded for each OTU (Table 3). Based on weighted pair group

cluster analysis, twelve association matrices (e. g., first generation,

Table 4; second generation, Table 5) were generated in the description

of a phenetic dendrogram (Figure 2).

The dendrogram in Figure 2 has utility in three ways:

1. It provides a visual illustration of the degree of taxonomic

similarity and difference between and among OTU's.

2. The taxonomic status of nominal genera may sometimes be re-

solved. The extreme resemblance of two genera, especially when one is

monotypic, may indicate a congeneric situation. This process has the

additional effect of providing a mechanism for evaluation and revision

of diagnostic characters.

3. It provides a basis for phylogenetic deductions, based on the

assumptions that a) phenetic clusters of extant OTU's are most likely

monophyletic, and b) the best estimate of the attributes of a common

ancestor of a cluster -- in the absence of direct evidence -- is pro-

vided by the cluster (Sokal and Sneath,.1963: 227).

Based on OTU association coefficients (Table 4, Figure 2) some

of the more apparent deductions about the phylogeny of Japanese carangids

include: "early" derivation of Chorinemus from the stem; derivation

of Trachinotus, and Elagatis, Naucrates, and Seriola from an "early"

common stem. These are not at variance with classical subfamilial

classification that relegates Chorinemus to the Chorineminae, Trachin-

otus, to Trachinotinae, and the latter three genera to the Naucratinae

(SUzuki, 1962). Were all lower clusters to be divided taxonomically,

however, it might be necessary to name the two remaining main stems

(which bear the nominal genera of the Caranginae). That is, two

groups of genera accorded to the subfamily Caranginae, namely Longirostrum

through Kaiwarinus, and Trachurus through Selar, plus Megalaspis

(Figure 2), deviate from a common stem at a lower coefficient than do

the Trachinotinae and the Naucratinae. Two alternatives to achieve

a uniform taxonomy are suggested: 1) assign subfamilial rank to the

two carangine groups, since Elagatis Seriola and Trachinotus are

accorded to subfamilies, despite a higher branch-point; or 2) abondon

the Naucratinae and Trachinotinae as subfamilial categories.

An additional conflict of taxonomies is seen for the Megalaspinae

(Megalaspis) which was thought by Suzuki to be an early offshoot of the

Naucratinae-Caranginae line (Figure 1). Cluster analysis of character

states reveals the affinity of Megalaspis to the Trachurus-Selar group

illustrated in Figure 2.

Agreement of the two taxonomies is also observed in the clustering

of the genera Atule, Decapterus, Trachurus, and Selar (=Trachurops).

The clustering of the genera of the Longirostrum Kaiwarinus group

also agrees with Suzuki's phylogeny.

In practice, the numerical taxonomy illustrated in Figure 2

points out the subjective nature of Suzuki's scheme. Based on ten

characters Suzuki (1962: 130-132) concluded that (1) the Naucratinae

are nearest the "ideal form" (the progenitor of the family), (2) the

Trachinotinae (more primitive than the Chorineminae) and the Chorineminae

both deviated at an early stage from the main stem leading to the

Carangidae, (3) the Megalaspinae are an offshoot from the line leading

the ancestral form of the Naucratinae to the Carangidae, and (4) the

evolution of the Caranginae represents "the main stem of the phylogene-

tical tree of the Carangidae." However, our lack of knowledge of

degree of primitiveness of many of the character states employed by

Suzuki reduces the confidence placed in his classification, although

the general affinities of many of the genera considered in both

studies reinforce many of the phylogenetic deductions to be drawn.

The phenetic dendrogram generated by cluster analysis in this

study avoids the speculative nature of differential character weighting

and its attendant difficulties (Sokal and Sneath, 1963: 118-120).

It provides an objective alternative for establishment of classifications,

with the utility of illustrating phylogenetic trends if the assumptions

noted above (3a and b) are valid. McAllister (1966: 234-235) mentioned

that perhaps the ideal procedure is the incorporation of objectively

weighted characters into such a scheme. The need for additional fossil

evidence as a basis for weighting the carangids is apparent, especially

as it pertains to the divergence of taxa from ancestral forms.


Infraspecific Variation

Examination of series of similar-sized individuals reveals that

infraspecific osteological variation is minimal in the genus Hemicaranx,

with the exception of the suborbital shelf. The shape of the suborbital

shelf is uniquely variable in Hemicaranx (Figure IV), and is unreliable

as a diagnostic character in distinguishing taxa. All other skeletal

features, however, are remarkably constant. Thus it is reasonable to

hypothesize that osteological features are relatively "conservative"

in the Carangidae. Consequently, previous descriptions and definitions

based on single specimens may be regarded as more valid and usually

accurate representations of the osteology of this group. In the compari-

son of caudal skeletons, for example, more confidence may be placed in

single-specimen observations of carangids. Indeed, this might be pre-

dicted on the basis of the observation of Schleuter and Thomerson (1971:

334) that little variation exists in the caudal skeleton of strong swim-

mers. However, the demonstration by Schleuter and Thomerson of signi-

ficant osteological variability in some other fishes cautions against

a typological approach to skeletal characters.

For the most part, ontogeny of the skeleton is characterized by a

uniform expansion of each element, thus allowing some comparison of

different-sized specimens. In the comparative section below, however,

individuals of nearly identical size always were compared.

Descriptive Osteology of Hemicaranx

Neurocranium (Figure I)

Prevomer (PV)(vomer of Suzuki, 1962: 48). -- The unpaired pre-

vomer is the anteriormost neurocranial bone. It is characterized in

the lateral plane by a forward-pointing triangular-shaped head from

which a median blade-like process extends posteriorly to articulate

with the ventral surface of the parasphenoid (PS). The anterior

margin of the head of the prevomer carries a dorsal crest, behind

which the ventral edge of the ethmoid (E) inserts. Just posterior

to this, the ventral edge of the lateral ethmoid (LE) articulates.

Antero-laterally the prevomer is in contact with the medial portion of

the palatine (PAL). Dentition is not present in individuals of 65

or 150 mm SL.

Ethmoid (E)(mesethmoid of Suzuki, 1962: 50). -- The ethmoid is

unpaired, and is compressed, vertically elongate, and narrowest at the

middle. The anterior ventral edge inserts behind the dorsal crest of

the prevomer (PV), while the posterior vertical edge is articulated

with the medial edge of the lateral ethmoid (LE). The dorsal posterior

corners of this bone are each overlaid by the pointed anterior tip of

the frontal bones (F). The dorsal surface is convex anteriorly.

The anterior edge of the dorsal surface of the ethmoid (Figure II)

is also convex in H. leucurus, but it is more sharply curved than in

zelotes. In both H. amblyrhynchus and bicolor it is concave, and is

characterized by projections on the anterior lateral corners.

Lateral ethmoid (LE)(ectethmoid of Suzuki, 1962: 50). -- The

lateral ethmoid is oriented in the vertical plane, and in lateral view

gently curves posteriorly upward from its articulation with the anterior

parasphenoid (PS). Viewed anteriorly, the lateral ethmoid appears broad

and somewhat butterfly-shaped. The medial edge of this bone broadly

articulates with the posterior lateral edge of the ethmoid (E), which

separates it from its fellow member. Dorsally the lateral ethmoid

contacts the anterior edge of the frontal (F). Located in the upper

medial quarter is the olfactory foramen. Ventrally, the lateral

ethmoid loosely contacts the anterior tip of the pterygoid bone (PT).

Frontal (F). -- The largest of the neurocranial elements, the

frontal bone is broadly united with its fellow member anteriorly in a

dorsally projecting median crest, which is an anterior continuation of

the supraoccipital crest. Anteriorly, the frontals are separated by

the interposed ethmoid (E), while posteriorly the supraoccipital (SUO)

is juxtaposed between them. The anterior lateral corner of the frontal

rests on the dorsal edge of the lateral ethmoid (LE). Together with the

parietal(P), to which it articulates posteriorly, the frontal contributes

to the dorsal-projecting temporal crest, which extends to the forward

corner of the frontal, terminating at the anterior margin of the orbit.

Laterally, a second crest, the pterotic crest, is present; anteriorly

it extends just in advance of the posterior margin of the orbit, and

posteriorly it carries on to the autopterotic bone (PTO), which articu-

lates with the posterior margin of the frontal. Beneath the temporal

crest, the ventral surface of the frontal articulates with the ptero-

sphenoid (PTS). Beneath the pterotic crest the ventral surface joins

the autosphenotic (SPH),

Pterosphenoid (PTS)(alisphenoid of Suzuki, 1962: 59). -- Hidden

from dorsal view, this small rhomboidal-shaped bone is widely separated

from its fellow member by the medial anterior opening of the braincase.

Dorsally the pterosphenoid articulates with the frontal (F). Posteriorly

it articulates with the autosphenotic (SPH); the ventral posterior angle

articulates with the prootic (PRO). The ventral edge of the ptero-

sphenoid joins the dorsal tip of the lateral wing of the basisphenoid

bone (BS).

Basisphenoid (BS). -- The unpaired basisphenoid is a Y-shaped,

short rod-like bone that runs vertically from near the posterior end

of the dorsal surface of the parasphenoid (PS) to join, via each short

arm, the pterosphenoids (PTS).

Parasphenoid (PS). -- The parasphenoid is unpaired and extends

along the midline of the floor of the orbit. Anteriorly, it is charac-

terized by a dorsal keel, and it is broadly joined to the dorsal surface

of the prevomer (PV) process. Laterally, the parasphenoid appears as

a long narrow shaft, from which an ascending process arises near the

posterior end. The ascending process flares away from its fellow as

it extends vertically to meet the lower anterior edge of the prootic

(PRO). The anterior tip of the parasphenoid touches the ventral margin

of the ethmoid (E). Posteriorly, the parasphenoid is broadly connected

to the anterior end of the basioccipital (BOC).

Supraoccipital (SUO). -- Prominent in lateral view is the medial

dorsal keel, the supraoccipital crest, formed by this unpaired bone.

The supraoccipital bone extends anteriorly over the posterior third of the

orbit. From front to back, it contacts the medial edges of the frontal (F),

parietal (P), and epiotic (EPO), respectively. Ventero-posteriorly it

joins the exoccipitals (EOC).

Parietal (P). -- The anterior edge of the parietal articulates

with the posterior edge of the frontal (F), from which it carries back-

ward the temporal crest. The parietal is widely separated from its

fellow member by the medially interposed supraoccipital (SUO). The

lateral edge articulates with the autopterotic bone (PTO).

Autopterotic (PTO)(pterotic of Suzuki, 1962: 60). -- The auto-

pterotic bone is prominently characterized by the pterotic crest, which

is carried forward by the adjoining frontal bone (F). Posteriorly the

crest terminates as a posterior projection, below which a spine extends

backward. The ventral edge is articulated with the autosphenotic (SPH)

anteriorly, and posteriorly with the prootic (PRO), the opisthotic (OPS),

and the exoccipital (EOC). Dorsally the pterotic contacts the parietal (P)

and the epiotic (EPO).

Autosphenotic (SPH)(sphenotic of Suzuki, 1962: 59). -- Characterized

by ridges and perforations, the autosphenotic is a small, strongly built

bone found at the posterior angle of the orbit. Anteriorly, the dorsal

edge unites with the ventral surface of the frontal beneath the pterotic

crest, and the medial edge is united with the lateral margin of the ptero-

sphenoid (PTS). The ventral surface of the autosphenotic articulates with

the prootic (PRO). Its posterior edge joins the autopterotic (PTO).

Epiotic (EPO). -- The pyramid-shaped epiotic bone supports a postero-

dorsal backward-projecting process to which the upper arm of the post-

temporal bone (PTM) articulates. The ventral corner of the epiotic is

united with the exoccipital (EOC). The antero-medial dorsal surface is

joined to the posterior edge of the supraoccipital (SUO). The antero-

lateral dorsal surface is joined to the dorsal posterior edge of the

parietal (P). Laterally it contacts the autopterotic (PTO).

Prootic (PRO). -- The irregularly hexagonal prootic bone is

prominent in the lower anterior lateral wall of the braincase. Dorsally,

the prootic is united with the pterosphenoid (PTS) anteriorly, the auto-

sphenotic (SPH), and posteriorly with the autopterotic (PTO). The

anterior edge of the prootic receives a portion of the arm of the basis-

phenoid (BS), and ventrally it also articulates with the ascending pro-

cess of the parasphenoid (PS). The ventral edge of the prootic approaches

and parallels the posterior end of the parasphenoid. Posteriorly,the

prootic is attached to the basioccipital (BOC), exoccipital (EOC), and

opisthotic (OPS).

Exoccipital (EOC). -- The exoccipital is located at the posterior

end of the cranium above the basioccipital (BOC), to which its ventero-

lateral margin is joined. Ventrally, the exoccipital meets its fellow

member to form the ventral margin of the foramen magnum. Each is dev-

eloped as a facet that articulates with the atlas vertebra. The post-

erior margins ascend to form the lateral margins of the foramen magnum.

The anterior edge of the exoccipital is joined to the prootic (PRO),

autopterotic (PTO), and opisthotic (OPS).

Basioccipital (BOC). -- The unpaired basioccipital is character-

ized by a concave, circular posterior end that articulates with the

anterior centrum of the atlas. Anteriorly, the basioccipital is strongly

united with the parasphenoid (PS), and it is connected to the posterior

margin of the prootic (PRO). Dorsally, it joins the lower edge of the

exoccipital (EOC).

Opisthotic (OPS). -- The opisthotic supports a posterior projection

that articulates with the lower arm of the posttemporal. Hidden from

dorsal and lateral view, it adjoins the autopterotic (PTO) and the exoc-

cipital (EOC) bones.

Suborbital series (Figure III). --

Lacrymal (L). -- This is the first, anteriormost element of the

suborbital series, and it sheaths the dorsal edge of the maxillary bone

(MX). The lacrymal is thin and flat, and extends from the anterior tip

of the maxillary back to its articulation with the second suborbital

bone (SUB) above the posterior margin of the palatine-pterygoid (PAL-PT)


Suborbitals (SUB). -- These four bones are linked to form the

posterior ventral quarter of the orbit. Prominent in dorsal view is the

suborbital shelf of the third suborbital bone (Figure IIIB). The fifth

suborbital is dorsally articulated with the dermosphenotic (DSO).

The suborbital shelf is unique in Hemicaranx in terms of its extreme

variability. Two examples observed in H. amblyrhynchus of 70 mm SL are

illustrated in Figure IV.

Dermosphenotic (DSO). -- The dermosphenotic resembles the sub-

orbitals and extends dorsally to cover the lower half of the autosphenotic

(SPH) bone.


Oromandibular region. --

Lower jaw bones (Figure V). --

Dentary (DN). -- The anterior end of the dentary meets its fellow

in a medial symphysis. From the anterior end two broad arms angle backwards.

The upper arm bears a single row of roundly pointed canine teeth, while

the lower arm articulates dorsally and posteriorly with the articular

bone (AR). Between the arms a notch receives and encloses the anterior

tip of the lower arm of the articular bone.

The posterior tip of the upper arm of the dentary is curved in H.

zelotes and bicolor, more pointed but still rounded in amblyrhynchus,

and sharply pointed in leucurus (Figure VI).

Articular (AR). -- Two arms extend forward from the base of the

articular. The upper tapers to a point; the lower is larger and pro-

ceeds horizontally with its ventral edge in close contact with the

dentary (DN) to enter into the notch of the dentary. The dorso-posterior

corner of the articular bears a facet which receives the ventrally

projecting knob of the quadrate (Q). The ventral posterior corner is

variably overlaid by the angular bone (AN).

Angular (AN). -- This small bone covers the posterior ventral

corner of the articular (AR), and is more readily seen in the medial


Upper jaw bones (Figure VII). --

Premaxillary (PMX). -- The anterior end of the premaxillary unites

with its fellow member to form the anterior margin of the upper jaw.

The ascending process of the premaxillary bone rides in a groove at the

anterior end of the maxillary (MX), and is longer than the articular

process. The articular process arises dorsally from the tooth-bearing

arm of the premaxillary and extends behind the middle of the maxillary.

Canine teeth are present in a single row along the ventral edge of the


The dorsal surface of the ascending process of the premaxillary

(Figure yVll) is indented in H. zelotes and bicolor, but in leucurus

and amblyrhunchus it is flattened. The posterior edge of the articular

surface of the premaxillary (Figure VIII) is broadly concave in leucurus

and amblyrhynchus, while in zelotes and bicolor it descends as a straight

edge from a dorsal extension.

Maxillary (MX). -- The maxillary is a slender bone, located above

and parallel to the length of the premaxillary (PMX). Anteriorly, the

head of the maxillary is grooved and articulates with the ascending pro-

cess of the premaxillary. Just behind the head, the maxillary fits snugly

into the ventral curve of the pre-palatine process of the palatine bone

(PAL). Posteriorly, the maxillary is expanded in the vertical plane, and

underlies the entire supramaxillary (SMX).

Supramaxillary (SMX). -- The supramaxillary is flat, pointed anter-

iorly, and meets the posterior end of the maxillary (MX) along its entire

ventral edge.

Hyoid region (Figure IX). --

Hyomandibular (HM). -- The hyomandibular is a stout, flattened bone

that is characterized by a descending hyomandibular process. Posteriorly

the hyomandibular articulates with the anterior preopercle (POP) margin.

Anteriorly, the hyomandibular process and the middle edge of the hyoman-

dibular are strongly joined to and partially underlain by the metaptery-

goid (MTPT). Dorsally the hyomandibular articulates with the neurocranium

at two points: an anterior knob articulates with the autosphenotic (SPH),

and the middle knob with the autopterotic (PTO). The posterior knob on

the head of the hyomandibular articulates with a facet on the upper

opercle ,(OP). The ventral tip of the hyomandibular process articulates

with the dorsal tip of the interhyal bone (IH).

Metapterygoid (MTPT). -- This bone articulates posteriorly with

the lower anterior edge of the hyomandibular (HM); together their

anterior margins curve parallel to the posterior suborbitals (SUB).

Anteriorly, the upper edge meets the posterior lateral margin of the

mesopterygoid (MSPT), while the lower edge articulates with the dorsal

rear margin of the quadrate (Q). The lower posterior margin of the

metapterygoid articulates with the upper posterior edge of the symplectic


Symplectic (SY). -- The symplectic is an elongate rod-like bone,

the anterior end of which firmly nestles in a groove on the medial sur-

face of the lower half of the quadrate (Q). From its firm union with

the quadrate, the symplectic runs posteriorly to articulate with the

ventral curve of the posterior metapterygoid (MTPT).

In lateral view the symplectic is characterized by variably developed,

mid-ventral expansions in all four species. In amblyrhynchus, bicolor,

and leucurus a dorsal expansion is also present. (Figure X).

Quadrate (Q). -- This small triangular bone is characterized by

an anterior ventral knob that articulates with a well-developed facet on

the articular bone (AR) of the lower jaw. The anterior edge of the

quadrate borders the posterior edge of the lower limb of the pterygoid

(PT). Posteriorly, the dorsal margin of this bone parallels but does

not contact the mesopterygoid (MSPT); the ventral margin projects

backwards and articulates with the inner curve of the lower limb of

the preopercle (POP). Medio-ventrally the quadrate firmly receives the

anterior symplectic (SY).

Mesopterygoid (MSPT). -- With the exception of a small downward tab

at the lateral margin, the mesopterygoid lies in a nearly horizontal

plane. It meets its fellow member medially below the entire length of

the parasphenoid (PS) to form a bony support for the roof of the mouth

and the floor of the orbits. Laterally, the downward tab separates the

metapterygoid (MTPT) and the pterygoid (PT). The anterior lateral edge

of the mesopterygoid articulates with the forward limb of the pterygoid;

its posterior lateral edge articulates with the anterior curve of the


Pterygoid (PT). -- The pterygoid bone is composed of two rod-like

arms that meet at an obtuse angle. The lower, downward-projecting arm

articulates along its posterior edge with the anterior margin of the

quadrate (Q). The upper, forward-projecting arm articulates along its

dorsal surface with the mesopterygoid (MSPT). Laterally, it articulates

with the posterior medial edge of the palatine (PAL). The anterior tip

of the pterygoid is in loose contact with the ventral edge of the lateral

ethmoid (LE) of the neurocranium.

The anterior end of the pterygoid bone (Figure XI) is developed as

a dorsally concave point in H. zelotes and leucurus. In amblyrhynchus

and bicolor it is also pointed, but is further characterized by indenta-

tions above and below the point.

Palatine (PAL). -- The palatine articulates with the anterior tip

of the lateral ethmoid (LE) via a medial, plate-like swelling, from which

two processes project. The posteriorward arm closely articulates with

the pterygoid (PT). The curved, anterior, laterally projecting arm (pre-

palatine process) articulates with the dorsal surface of the anterior end

of the maxillary (MX). The palatine also contacts the neurocranium by

riding the anterior lateral surface of the prevomer (PV).

Opercular bones. -- The four opercular bones form the gill cover

and are variously connected with each other and the hyomandibular series.

Opercle (OP) (Figure XIIA). -- The opercle is a large flattened

bone, curved broadly posteriorly, that articulates with the posterior

knob of the hyomandibular (HM) via a facet at its antero-dorsal corner.

The anterior margin underlies the posterior margin of the preopercle (POP).

Ventrally the opercle covers the dorsal margin of the subopercle (SOP).

Subopercle (SOP) (Figure XIIA). -- The subopercle is flattened and

is characterized by a tapered forward-projecting process that arises from

the lower anterior corner. The process and the leading edge of the sub-

opercle are covered by the ventral tip of the opercle bone (OP). The

lower anterior corner is covered by the posterior edge of the interopercle


Interopercle (IOP) (Figure XIIA). -- In addition to covering the

anterior corner of the subopercle (SOP), the interopercle articulates

mid-dorsally with the epihyal bone (EH). The interopercle is in turn

covered along its dorsal half by the ventral margin of the preopercle

(POP). The anterior end of the interopercle is linked by strong connect-

ing tissues to the posterior margin of the mandible.

Preopercle (POP)(Figure XIIB). -- The anterior margin of the upper

limb of the preopercle is firmly joined with the posterior edge of the

hyomandibular bone (HM). Posteriorly,it covers the dorsal half of the

interopercle (IOP) and the anterior edge of the opercle (OP). The dorsal

surface of the lower, forward-projecting limb articulates with the ventral

edge of the posterior projection of the quadrate (Q). Juvenile specimens

are characterized by preopercular spines (Figure B) that radiate poster-

iorly from the posteo-ventral angle of the preopercle.

Hyal bones (Figure XIII and XIV). --

Basihyal (BH) (glossohyal of Suzuki, 1962: 89). -- This anterior-

most of the hyal series is an elongate, rod-like,unpaired bone that forms

the base of the tongue. Its posterior end fits snugly into a rounded

notch formed by the convergence of the first basibranchial (BB) and the

upper hypohyals (UHH) (Figures XIII and XIV).

The basihyal (Figure XVB) is anteriorly widest in H. zelotes, inter-

mediate in leucurus and amblyrhynchus, and narrowest in bicolor. Pointed

conical teeth are found on the dorsal surface of the basihyal in

amblyrhynchus and bicolor.

Upper hypohyal (UHH). -- This curved, trapezoidal bone, which ap-

proaches its fellow member at the medial tip of the dorsal surface, is

loosely articulated with the first basibranchial (BB) (Figure XIII).

In concert the three form a concave notch which receives the posterior end

of the basihyal (BH). The ventral edge of the upper hypohyal parallels

but does not touch the dorsal edge of the lower hypohyal (LHH). Posteriorly,

the upper hypohyal articulates with the upper anterior margin of the

ceratohyal (CH). A circular foramen pierces the upper posterior quadrant

of this double-layered bone.

In lateral view, the foramen located in the upper posterior quarter

of the upper hypohyal bone (Figure XVC) is variably shaped, being elongate

in zelotes and leucurus and more circular in amblyrhynchus and bicolor.

Lower hypohyal (LHH). -- The lower hypohyal curves inward medially

to meet its fellow member ventrally. Dorsally, it is loosely aligned with

the upper hypohyal (UHH). Posteriorly, it receives an underlying pro-

jection from the anterior edge of the ceratohyal (CH) to which it is

broadly articulated.

Ceratohyal (CH). -- Prominent in the upper central area of this

elongate bone is the oval, ventrally-skewed ceratohyal window. The

ceratohyal bone is closely articulated with the lower hypohyal (LHH)

via an anterior projection arising from its lower forward edge. The

upper anterior edge articulates with the posterior edge of the upper

hypohyal (UHH). The ceratohyal is strongly united posteriorly with the

epihyal bone (EH). Ventrally, it receives the proximal tips of five

branchiostegal rays (B).

The ceratohyal window is ventrally-expanded in H. zelotes and

amblyrhynchus, more oval in bicolor, and roughly triangular in leucurus

(Figure XVD).

Epihyal (EH). -- The epihyal is a rounded triangular-shaped bone

that is strongly united anteriorly with the ceratohyal (CH). Postero-

ventrally it receives the proximal tips of the last two branchiostegal

rays (B). The mid-upper lateral surface of the epihyal articulates

with the mid-dorsal edge of the interopercle (1OP). The upper posterior

angle of this bone is a facet that receives the ventral end of the

interhyal bone (IH).

Urohyal (UH). -- The urohyal is an unpaired rectangular bone that

lies in the ventral vertical plane and is characterized by a pair of

laterally flared wings at the ventral edge. Anteriorly,the urohyal is

loosely inserted -into a space surrounded by the first basibranchial (BB),

the hypohyals (HH), and the anterior portion of the ceratohyals (CH).

The urohyal (Figure XVA) of H. zelotes is uniquely characterized

by slight posteriorward spines on the dorsal edge. In addition, a pro-

jection at the ventral anterior corner is variably developed, being

longest in bicolor and shortest in zelotes.

Interhyal (IH). -- The interhyal is a short rod-shaped bone that

links the hyal series to the hyomandibular bone (HM). Dorsally, it

articulates with the hyomandibular; ventrally it meets a facet on the

posterior dorsal corner of the epihyal (EH).

Branchiostegal rays (B). -- Seven branchiostegal rays proceed

posteriorly from the hyal series to lend support to the lower opercular

membrane. The five anteriormost branchiostegals arise from the cera-

tohyal (CH); the last two of these lepidotrich-like rays arise from the

epihyal (EH).

Branchial region. -- Five branchial arches composed of paired and

unpaired elements are found in Hemicaranx (Figure XIII).

Basibranchials (BB). -- Three unpaired basibranchial bones are

located in a longitudinal series in the floor of the pharynx and re-

present the ventralmost members of the branchial series. The first is

only partially visible in dorsal view since it is overlaid anteriorly

by the basihyal bone (BH) (Figure XIII), below which it vertically extends

(Figure XIVB). In concert with the upper hypohyals (UHH) the first

basibranchial forms a concavity that receives the posterior end of the

basihyal, and thus links the branchial and hyal series. Posteriorly,

the first basibranchial articulates with the second basibranchial.

The second basibranchial bone is short, rod-like, and from its

articulation with the first basibranchial extends dorso-posteriorly

to meet the third basibranchial. The second basibranchial is located

slightly above the dorsal anterior tips of the paired first hypo-

branchials (HB), with which its lateral anterior end articulates. The

posterior corners cover, but do not articulate with, the proximal ends

of the second hypobranchials.

The elongate third basibranchial extends obliquely back from the

articulation of its concave anterior end with the concave posterior

tip of the second basibranchial (Figure XIVB). The anterior lateral

margins of this bone firmly receive the dorsal halves of the proximal

ends of the paired second hypobranchial (HB). Posteriorly, the third

basibranchial is loosely cradled by the dorsal ends of the third hypo-


The third basibranchial is slightly more expanded laterally in H.

leucurus and bicolor than in zelotes and amblyrhynchus.

Hypobranchials (HB) (Figure XIII). -- Three pairs of hypobranchials

are variously articulated with the basibranchial bones (BB) and extend

laterally to articulate with the first three ceratobranchials (CB).

The first and anteriormost hypobranchial bone is compressed and

firmly links the first ceratobranchial to the second basibranchial,

which possesses a shallow recess on the anterior lateral surface into

which the upper half of the proximal end of the hypobranchial fits.

The ventral edge of this elongate bone supports gills, while the

lateral and medial sides support gill rakers and tubercles, res-


The second hypobranchial is more rectangular than the first, with

gills, gill rakers and tubercles similarly supported. The dorsal half

of the proximal end inserts into a recess along the anterior third of

the lateral margin of the third basibranchial.

The third hypobranchial is roughly funnel-shaped in appearance,

with the narrow tube-like process extending down and forward to ap-

proach its fellow member in the midline below the middle of the third

basibranchial. The medial margin of the flared dorsal end of this

bone flanks the posterior end of the third basibranchial. The posterior

end supports the third ceratobranchial.

Ceratobranchial (CB)(Figure XIII). -- The first four are gently

curved slender rods that dorsally support epibranchials (EB), and with

the exception of the fourth, are supported ventrally by the hypobranchials

(HB). Except for the first, which supports long gill rakers instead of

tubercles on its lateral margin, each supports two rows of tubercles on

the dorsal surface.

The fifth ceratobranchial (or pharyngeall") is more massive and

laterally expanded than the others, and is densely covered dorsally by

pointed slender projections (teeth). Like the fourth ceratobranchial,

this element is not articulated with a hypobranchial; both are connected

to the base of the hyal series by cartilage. The fifth ceratobranchial

is also securely joined to its fellow member along the anterior medial


Epibranchials (EB)(Figure XIII). -- The four stout epibranchial

bones effect the characteristic sharp bend of the gill arches and link

the first four ceratobranchials (CB) to the pharyngobranchial series.

Each is laterally compressed and characterized by a variably developed,

wing-like, posterior projection.

Pharyngobranchials (PB)(Figure XIII). -- The first pharyngobranchial

bone is a slender rod that connects the first epibranchial bone (EB) to

the neurocranium. The other three pharyngobranchials are solidly built

blocks of bone characterized by long, slender sharp teeth projecting

into the pharyngeal cavity, and they are connected to the neurocranium

by a strong combination of muscle and connective tissue. The third is

the largest.

Appendicular Skeleton

Pectoral girdle and fin (Figure XVIA). --

Posttemporal (PTM). -- This anteriorly bifurcated bone supports all

other pectoral elements by its articulations with the neurocranium and

the supracleithrum. The upper arm of the posttemporal articulates with

the epiotic process, and the lower arm articulates with the opisthotic

(0PS) of the skull. The supracleithrum (SCL) meets it postero-ventrally.

Supracleithrum (SCL)(supraclavicle of Suzuki, 1962: 108). -- This

flat, elongate bone extends obliquely downward from its dorsal articula-

tion with the posttemporal (PTM). Ventrally it is attached to the lateral

upper angle of the cleithrum (CL).

Cleithrum (CL)(clavicle of Suzuki, 1962: 109). -- The cleithrum

is an elongate, broadly curved bone that is characterized ventrally by

two posteriorly projecting shelves, the exterior and interior, that meet to

form a ridge anteriorly. Dorsally, a broad shelf extends back from the

curved main axis of the bone. The cleithrum articulates dorso-laterally

with the supracleithrum (SCL). Medially, it touches the dorsal tip of

the postcleithrum (PCL) and borders the lateral dorso-anterior edge of

the scapula. Ventrally, the cleithrum touches its opposite member in

the midline of the body.

Scapula (SC)(hypercoracoid of Suzuki, 1962: 110). -- The scapula

is small, flat, and roughly rhomboidal in shape, with a central foramen.

Located behind the middle of the cleithrum (CL), this bone provides sup-

port for the pectoral spine and at least the upper three radials (R).

It articulates anteriorly with the medial surface of the cleithrum, and

its entire ventral edge borders the dorsal edge of the coracoid (CO).

Coracoid (CO)(hypocoracoid of Suzuki, 1962: 110). -- This thin,

slightly concave bone is closely bordered dorsally by the scapula (SC),

and is characterized by a concave facet at the posterior corner of this

articulation. Anteriorly, the upper and lower edges of the coracoid

approach the posterior margin of the internal shelf of the cleithrum


Postcleithrum (PCL)(postclavicle of Suzuki, 1962: 112). -- The

postcleithrum is actually composed of two broadly overlapping bones.

The upper element (PCL 1) is elongate and expanded ventrally, and

articulates dorsally with the medial upper surface of the posteriorly

extending process of the cleithrum (CL). Ventrally, it extends to a

point even with the lowest radial (R). The lower thin, rib-like element

(PCL 2) firmly articulates with the upper element as far as the lower

edge of the posterior cleithral process.

Radials (R). -- These four stout rods provide support for the soft

fin rays. The upper three are, in turn, supported by the posterior edge

of the scapula (SC), and the fourth loosely articulates with the coracoid


Lepidotrichs (LEP). -- The lepidotrichs are represented by one fin

spine, plus 18 23 soft rays (Table 17), all but the first of which are

branched. The spine articulates directly with the scapula (SC); the

soft rays are supported by the radials (R), which they fringe in a broad


Pelvic girdle and fin (Figure XVIB).---

Basipterygium (BPT). -- The basipterygium is closely aligned with

its fellow member along its entire medial axis, tapering slightly from

a knobby posterior to a splint-like anterior that passes medial to the

coracoid (CO) and cleithrum (CL), and below the postcleithrum (PCL).

From the medial corner of the posterior end arises a splint-like process

that projects obliquely upward. In lateral view the splint appears to

carry through the main column of the bone to project ventrally. Six

lepidotrichs (LEP) are supported by the posterior margin of the


Lepidotrichs. -- One spine at the lateral corner, and five soft

rays medial to it, are supported by the posterior basipterygium (BPT).

Postcranial Axial and Medial Skeleton

Dorsal fin. Two distinct parts of the dorsal fin are present in

adult and advanced juvenile specimens of Hemicaranx. The structural

continuity of the dorsal fins is illustrated by the regular spacing of

the lepidotrich-supporting pterygiophores (PT) (Figure XVII).

Anteriormost in the dorsal skeleton are three predorsal bones (PD),

the first placed before the first neural spine (NS), the second and

third lying between the second and third neural spines. Following

the predorsals are pterygiophores (PT), which underlie and support

the lepidotrichs (LEP). With the exception of the first member, each

pterygiophore is segmentally associated with a lepidotrich and also

underlies structurally the lepidotrich immediately preceding its

segmental associate. All but the first and last lepidotrich have

compound support. (According to Smith and Bailey [1961: 348], the

pterygiophore of the first dorsal spine is actually included in the

predorsal series.) The one-to-one correspondence is illustrated in

Figure XVII. The true nature of lepidotrich support is shown in

Figure XVII. As seen in Figure XVIIA, dorsal spines II through IX

are segmentally supported by an elongate proximal pterygiophore (PPT)

and a short posteriorly-pointed distal pterygiophore (DPT). All soft

rays are supported by an elongate proximal pterygiophore and a bi-

laterally halved distal pterygiophore inserted between halves of the

segmentally associated lepidotrich (Figure XVIIIB). The trend of

fusion of anterior proximal and intermediate pterygiophores in spiny-

rayed fishes (Smith and Bailey, 1961: 347) is completed in H. zelotes;

in the only other comparably sized specimens (SL: 33 mm) of Carangidae

discussed in the literature, Elagatis bipinnulata has only the two

posterior intermediate pterygiophores distinct (Berry, 1969: 456-457).

Insertion of proximal pterygiophores between neural spines is shown

in Figure XVII.

Anal fin (Figure XVII). -- Structurally and ontogenetically the

anal fin resembles the dorsal fin with regard to overall appearance,

and pterygiophore lepidotrich structure and articulation. Unique

is the anteriormost proximal pterygiophore (PPT), which articulates

with the hemal spine (HS) of the eleventh vertebra to form a strong

brace-like structure. It is also extended forward at its ventral base.

The distal tips of the proximal pterygiophores of each species

are characteristic in development of anterior-directed points and

degree of opposition to each other (Figure XIX). The pterygial

points are most prominent in H. zelotes and are developed on all but

the anterior two normal pterygiophores. They are poorly developed in

leucurus and are pointed only on the tenth through the eighteenth

normal pterygiophores. Points are intermediately developed in

amblyrhynchus and bicolor. (Points on dorsal pterygiophores follow

a similar pattern, but they more closely resemble each other.)

The distal ends of the anal pterygiophores are most closely op-

posed in H. amblyrhynchus, intermediately so in bicolor and leucurus,

and most distantly articulated in zelotes. (Figure XIX)

Vertebral column. -- The vertebral column is composed of 25

vertebrae, all of which typically feature processes for articulation

with adjacent members and a neural arch and spine. The ten precaudal

vertebrae (PCV).possess pleural (PR) and/or epipleural (EPR) ribs;

hemal spines (HS) are variably developed. Fifteen caudal vertebrae

(CV) possess closed hemal arches and hemal spines that descend to

interdigitate with anal pterygiophores.

The first vertebra, the atlas (Figure XXA), is uniquely characterized

by a neural spine (NS) that is autogenous; it articulates with paired

facets on the atlas by two projections that jut inward from laterally

flared processes. Anteriorly the atlas articulates with the neurocranial

exoccipital (EOC) and basioccipital (BOC) bones. Posteriorly projecting

lateral apophyses extend past the articulation with the second vertebra.

The second vertebra, the axis (Figure XXA), is characterized by

neural prezygapophyses that extend over the atlas. Lateral apophyses

project backwards.

Vertebrae three through ten closely resemble each other, varying in

the degree of development of the features they share. As illustrated in

Figure XVII, neural postzygapophyses closely articulate with the follow-

ing neural prezygapophyses, which are successively more elongate.

Similarly, the lateral apophyses lengthen posteriorward, and from the

seventh on are pierced by a foramen. Commencing with the eighth vertebra,

hemal prezygapophyses are developed.

The eleventh vertebra, which is characterized by a hemal spine

broadly united with the first anal pterygiophore, and all other caudal

vertebrae back to the antepenultimate are characterized by well-

developed neural and hemal pre and postzygapophyses (Figure XXB).

(The terminal three vertebrae are discussed as part of the caudal


The pleural ribs articulate variously with the vertebrae, with

the third through the eighth inserting into hollow depressions at

the dorsal posterior angle of the lateral apophysis, and the others

articulating with the lateral vertebral surface. Epipleural ribs

articulate either directly with the vertebrae or with the proximal

end of a pleural rib, and are associated with vertebrae 1 through 14.

Numerically, the axial and medial skeletons of H. zelotes are

summarized as follows: predorsal formula: 0-0-0-2; dorsal softrays,

24 31; anal softrays 22 25; number of pterygiophores one less

than lepidotrichs of respective fins; vertebrae: 10 + 15.

In H. leucurus there are also 10 + 15 vertebrae, while 10 + 16

vertebrae are present in amblyrhynchus and bicolor. (Dorsal and anal

ray counts are summarized in Tables 15 and 16, respectively.)

Caudal Skeleton

The caudal skeleton of H. zelotes is formed by the terminal three

caudal vertebrae (CV), modified and unmodified neural and hemal spines,

and lepidotrichs (LEP) that articulate with or oppose the outer margin

of the bony caudal complex (Figure XXI). Anteriormost in the caudal

skeleton is the thirteenth caudal centrum, the spool-shaped ante-

penultimate (CAP). The neural spine (NS) of the antepenultimate

centrum extends posteriorly to underlie the anteriormost superior

secondary caudal rays (SCR), and the autogenous hemal spine (HS)

underlies all but the last two inferior secondary caudal rays. The

neural spine of the penultimate centrum (CP) is reduced posteriorly

and occupies the semi-circular space formed by the opposing edges of

the first epural and the uroneural bone. The autogenous hemal spines

of this centrum extends downward to underlie the two posteriormost

inferior secondary caudal rays. The urostylar vertebra (UR) is

formed anteriorly by a typical spool-shaped centrum; its posterior.

portion js reduced to an obliquely ascending projection. This uro-

stylar projection is ultimately fused posteriorly with the dorsal-

most hypural element, and dorsally with the paired uroneurals. It

articulates posteriorly with the free remaining hypurals.

One pair of uroneural bones (UN) are present in zelotes. They

are roughly "L"-shaped, with the anteriorward short arm fused to the

dorsal margin of the urostylar centrum, and the long arm fused to

almost the entire length of the dorsal-most hypural (Figure XXIIA).

Two epural bones (EP) articulate with the dorsal surface of the uro-

neurals. The first, which fits between the uroneurals with a rounded

ventral projection, extends anteriorly over the posterior half of the

penultimate neural spine; it also projects posteriorly to support sev-

eral superior secondary caudal rays. The second epural is a rod-like

bone that parallels the first and articulates with the uroneurals and

the posterior two secondary caudal rays.

The hypural plate of H. zelotes is formed of two hypurals (HY)

that form the inferior portion, and two hypurals in the superior half.

A fifth hypural dorsally is fused to the uroneurals longitudinally, and

to the urostylar vertebra at its proximal tip. Articulating with the

distal margins of the hypurals are the branched principal caudal rays


Nine superior plus eight inferior principal caudal rays are

present: all but the dorsal and ventral-most are branched. Additionally,

nine superior and eight inferior secondary caudal rays are present.

Ontogeny of the caudal skeleton of Hemicaranx zelotes is charac-

terized by fusion of certain elements, and this is representative of

the percoid fishes and their derivatives (Gosline, 1961: 268). As

noted in Figure XXIIA, fusion of the paired uroneurals to the dorsal

urostylar surface and to the dorsal-most hypural bone is completed by

the 33 mm stage. At this stage in development lines of fusion are still

apparent, but soon thereafter the complex appears as one bone (Figure

XXIIA). Although an earlier stage of zelotes was not available for

study, it is likely that no elements are masked at the 33 mm stage.

That is, lines of fusion are still apparent, although they quickly

become lost with growth. An additional fusion is noted between the

third and fourth hypurals (Figure XXII), but at 65 mm SL no indication

of their union remains. Ontogeny resembles that of Chloroscombrus

chrysurus; fusion of the same bones occurs. It is probable that

Figure XXIIB illustrates the sequence of fusion in zelotes.


IA. Body deeper, 40-50% SL; upper caudal fin lobe longer, 30-45%

SL; prevomerine dentition absent; interorbital width wider,

9-14% SL; pectoral fin 30-40% SL; exoccipital zygapophyses

adjoined; cranial depth exceeds width.

...Hemicaranx ... 2, p. 50

IB. Body shallower, 28-39% SL; upper caudal lobe shorter than one-

third SL; prevomerine dentition present; interorbital narrow,

7-10% SL; pectoral fin less than one-third SL; exoccipital

zygapophyses separate; cranial width exceeds depth.

...Atule ........ 5, p. 130

2A. Spinous dorsal rays seven; ratio of straight portion of

lateral line to curved portion of lateral line = 2.3-3.0;

proportional width of anterior caudal peduncle scute

exceeds 38 thousandths SL; caudal vertebrae 16; upper hypo-

hyal window circular; anterior end of pterygoid indented.

.... 3

2B. Spinous dorsal rays eight; ratio of straight portion of

lateral line to curved portion of lateral line = 1.7-2.3;

proportional width of anterior caudal peduncle scute width

less than 36 thousandths SL; caudal vertebrae 15; upper

hypohyal window oval; anterior end of pterygoid concave.

3A. Dorsal soft rays usually 27 or 28, range 24-30; anal soft


rays usually 24, (frequently 23 or 25), range 21-25;

upper caudal fin lobe relatively long, exceeding lower

lobe; anal pterygiophores closely spaced; ceratohyal

window intermediate, neither oval nor triangular.

...H. amblyrhynchus... p.63

38. Dorsal soft rays usually 25 or 26, range 24-28; anal

soft rays usually 22, (frequently 21 or 23), range 21-24;

upper caudal lobe equal to lower lobe length;.anal

pterygiophores more distantly spaced; ceratohyal window


...H. bicolor ........ p.93

4A. Caniform teeth roundly pointed; pectoral fin relatively

short, 30% SL in largest specimens; anteriormost caudal

peduncle scute wider, 2.6-3.6% SL; four to six lateral

body bars, fading with age; ceratohyal window intermed-

iately shaped, neither triangular nor oval; urohyal spine

present; anal pterygiophore points prominent and distantly


...H. zelotes ........ p.99

4B. Caniform teeth sharply pointed; pectoral fin long, more

pronounced with age, almost 40% SL in largest specimens;

anteriormost caudal peduncle scute narrower, 1.7-2.3% SL;

six to nine lateral body bars, fading with age; ceratohyal

window triangular; urohyal spine absent; anal pterygiophore

points reduced and intermediately spaced.

...H. leucurus ....... p.123

5A. Number of scutes in lateral line 33 to 49, usually 36 to 39.


.... 6

58. Number of scutes in lateral line 48 to 63, usually 53 to 55.


6A. Straight/curved ratio of lateral line sections ca. 2.0;

premaxillary dentition uniseriate.

...A. djedaba ........ p.163

6B. Straight/curved ratio of lateral line sections ca. 1.5;

premaxillary dentition biseriate anteriorly.

.... 7

7A. Ventral outline strongly convex; body deep, 35-40% SL;

anterior caudal peduncle scute wide, ca. 48 thousandths

SL; dentary teeth triseriate posteriorly; ultimate dorsal

and anal rays approximately equal to penultimate rays.

...A. kalla .......... p.156

78. Ventral outline no more curved than dorsal outline;

body shallower, ca. 30% SL; anterior caudal peduncle

scute narrow, ca. 36 thousandths SL; dentary teeth

uniseriate posteriorly; ultimate dorsal and anal rays

relatively longer than penultimate rays.

...A. mate ........... p.150

8A. Spinous dorsal blackened; supramaxillary not extended

forward as a point; anal rays 19 to 21; curved lateral

line scales 35-45.

...A. malam .......... p.167

8B. Spinous dorsal not blackened; supramaxillary extended

forward as a point; anal rays 21 to 23; curved lateral

line scales 41-55.
...A. macrurus ....... p.160


Hemicaranx Bleeker

Hemicaranx Bleeker, 1862: 135-136 (type species Hemicaranx marginatus

[=Hemicaranx bicolor], by original designation).

Carangops Gill, 1862: 435 (type species Caranx heteropygus [=Hemicaranx

amblyrhynchus], by original designation).


Carangid fishes characterized by uniseriate dentition on premaxil-

lary and dentary bones; jaw teeth fine, conical, pointed; depth 40 to

50% of standard length; lateral line with posterior-pointing scutes on

straight portion; no scutes on lateral line arch; no caudal peduncle

keels; premaxillary narrow, less than orbit; no dentition on prevomer;

ethmoid-prevomer keel elevated; exoccipital zygapophyses adjoined;

cranial depth greater than width; olfactory cavity well-developed;

propercular width more than one-third of preopercular length; cerato-

hyal window wide and ovoidal; caudal vertebrae 15 or 16.

Hemicaranx is distinguished from its most closely related carangid

genus, Atule, in having adjoined exoccipital zygapophyses, an elevated

ethmoid-prevomer keel, no prevomerine dentition, a distinct myodome

opening, cranial depth greater than width, body deeper (40-50% SL),

longer upper caudal fin lobe (30-45% SL), slightly longer pectoral

fin (30-40% SL), greater interorbital width (9-14% SL).

Hemicaranx is distinguished from Selar by having adjoined exoc-

cipital zygapophyses, no prevomerine dentition, a pterotic window,


cranial depth greater than width, the ascending process of the pre-

maxillary longer than the articular process, the ceratohyal window

wide and ovoidal, and the absence of two papillae on the shoulder


In addition to the characters that distinguish it from Selar,

Hemicaranx differs from Decapterus and Trachurus by a mesopterygoid

bone that is less than one-half the hyomandibular. Hemicaranx is

further distinguished from Decapterus by a urohyal shorter than the

hyoid body, a 90-degree angle of the cleithrum shelf, and the absence

of detached finlets posterior to the soft median fins; in addition it

differs from Trachurus in having the posttemporal height less than one-

half its length and the anterior scales in the lateral line not trans-

versely expanded.

From the genera Carangoides, Gnathanodon, Longirostrum, Selaroides,

and Uraspis, which apparently comprise a natural cluster (Figures 2 and 3),

Hemicaranx is distinguished by a low frontal-supraoccipital crest,

ceratohyal window wide and ovoidal, and more than 14 caudal vertebrae.

It differs from all genera except Selaroides in having the posttemporal

height less than one-half the length. It is distinguished from all

genera but Longirostrum by a preopercular width less than one-third the

width and more than 14 caudal vertebrae. Hemicaranx differs from

Gnathanodon and Selaroides in not having a well-developed metapterygoid

lamina and these three genera differ from the remaining three genera

(Uraspis, Carangoides, Longirostrum) in lacking prevomerine dentition.

Hemicaranx also differs from Carangoides and Uraspis in possessing a

distinct myodome opening. Hemicaranx is further distinguished from

Uraspis by its well-developed olfactory cavity, elevated ethmoid-prevomer

keel, pointed scutes directed forward, and lack of milky white areas

in the mouth; from Selaroides by a cranial depth greater than width,

and an ovoidal rostral cartilage; from Longirostrum by adjoined exoc-

cipital zygapophyses; from Gnathanodon by a dorsal process of the arti-

cular bone less than one-half its height; and from Carangoides by uni-

seriate dentition in both jaws.

Hemicaranx is distinguished from Chloroscombrus by the absence of

prevomerine dentition, a pterotic window, cranial depth greater than

width, opercular length less than interopercle length, posttemporal

ventral branch not attached to upper opisthotic, posttemporal height

less than one-half its length, caudal vertebrae more than 14, lower

jaw teeth uniseriate, and a ventral outline that is not appreciably

more convex than the dorsal outline.

From the remaining Caranginae genera, namely Alectis, Atropus,

Caranx, Citula and Kaiwarinus, which as a phenetic group Hemicaranx

less closely resembles, Hemicaranx is distinguished by a low frontal-

supraoccipital crest, an elevated ethmoid-prevomer keel, preopercular

width less than one-third its length, opercular length less than inter-

opercle length, posttemporal height less than one-half its length,

more than 14 caudal vertebrae, and upper jaw teeth uniseriate. Hemicaranx

is further distinguished from all genera except Alectis, by the absence

of prevomerine dentition. It shares, along with Alectis and Citula, a

pterotic window; and with Alectis and Kaiwarinus it shares a well-develop-

ed olfactory cavity. Hemicaranx and Kaiwarinus are distinguished by a

pre-palatine process directed laterally, and a wide and ovoidal cerato-

hyal window; these two genera, along with Atropus, are characterized

by a pterygoid bone that is neither enlarged nor acuminate. Hemicaranx

is further distinguished from Atropus and Caranx, which have a round

postmaxi,llary process, and from Alectis and Citula, with opercular

apparatus height exceeding length and pluriseriate upper jaw dentition.

Hemicaranx also differs from Citula in having a conspicuous pterotic

crest that is produced backward.

Hemicaranx differs from Megalaspis in having exoccipital zygapo-

physes adjoined, a well-developed olfactory cavity, an elevated ethmoid-

prevomer keel, a pterotic window, a distinct myodome opening, cranial

depth greater than width, ceratohyal window wide and ovoidal, caudal

vertebrae more than 14, prevomerine dentition absent, postmaxillary

process not rounded, and uniseriate upper jaw teeth.

Hemicaranx is distinguished from Elagatis, Naucrates, and Seriola.

by a well-developed olfactory cavity, an elevated ethmoid-prevomer keel,

no prevomerine dentition, cranial depth greater than width, post-maxil-

lary process triangular, ceratohyal window wide and ovoidal, postcora-

cold process undeveloped, ventral element of postcleithrum rib-like,

lateral-line scutes, first hemal spine enlarged, and lower jaw teeth

uniseriate. Hemicaranx and Seriola are both characterized by a pterotic

window and the mesopterygoid less than one-half the hyomandibular; with

Naucrates they are distinguished from Elagatis by a preopercular width

less than one-third the height and an opercular apparatus that is taller

than wide. Seriola is unique in having the opercular length exceeding

the interopercle length.

Hemicaranx is distinguished from Trachinotus by a low frontal-

supraoccipital crest, absence of prevomerine dentition, a triangular

postmaxillary process, uniseriate dentition, a well-developed olfactory

cavity, ethmoid-prevomer keel elevated, a pterotic window, myodome

opening distinct, a supramaxillary, metapterygoid less than one-half

the hyomandibular, ceratohyal window wide and ovoidal, posttemporal

ventral branch elongate, posttemporal height less than one-half its

length, lateral line scutes, and more than 14 caudal vertebrae.

Hemicaranx is most distantly related to Chorinemus, from which it

is distinguished by a well-developed olfactory cavity, ethmoid-prevomer

keel elevated, pterotic crest conspicuous and produced backwards, a

pterotic window, premaxillary protractile, a dentary-articular interstice,

mesopterygoid less than one-half hyomandibular, preopercular width less

than one-third its length, pre-palatine process directed antero-ventrally,

height of opercular apparatus exceeding its length, seven branchiostegal

rays, ceratohyal window wide and ovoidal, posttemporal height less than

one-half its length, and lateral line scutes. Chorinemus is further

characterized by prevomerine dentition, an extremely long maxillary,

maxillary length much greater than height, posttemporal ventral branch

attached to upper opisthotic, and pluriseriate dentition.


Morphometric measurements are summarized in Tables 7, 9, 11, 13;

meristic counts appear in Tables 8, 10, 12, 14, 15, 16, 17, 18, 19;

osteology is completely described in the osteological section.

Body compressed, but neither unusually depressed nor elongate;

body depth 40 to 50% of standard length; total length of lateral line

75% SL; lateral line dorsally arched anteriorly, becoming straight be-

neath anterior rays of soft dorsal fin and then continuing posteriorly

onto the caudal peduncle; 30 to 40 scales in lateral line arch, the arc

of which is roughly half the straight lateral line distance, and which

is in turn covered by 40 to 50 scutes; arch three times as long as high;

caudal peduncle slender, as wide as deep; dorsal outline of body broadly

convex, with soft dorsal origin midway between snout and hypural base;

soft dorsal and anal fins long and low, partially sheathed at base by

scaled membrane; upper caudal fin lobe 30 to 45% SL; lower caudal fin

lobe 30 to 33% SL; posterior margins of caudal lobes forming a broad,

slightly obtuse angle; origin of soft anal fin slightly behind origin

of soft dorsal fin, both extending to anterior part of peduncle; pectoral

fin 30 to 40% SL; pelvic fin 10 to 13% SL; pelvic fins inserted ventrally

just behind laterally-inserted pectoral fins; pectoral and caudal lobes

becoming pointed with age; head length 25 to 30% SL; interorbital width

9 to 14% SL; snout blunt, one fourth of head length, which is exceeded

slightly by head depth; orbit barely longer than snout; posterior end of

upper jaw terminating below anterior margin of orbit; maxillary depth

1.7 to 2.7% SL; gape 7 to 8% SL; teeth uniseriate and caniform in pre-

maxillary and dentary; teeth absent from roof of mouth; dorsal spines

seven or eight; dorsal soft rays 24 to 31; anal soft rays 20 to 25; pector-

al rays 18 to 23; pelvic rays 5; principal caudal rays 9 + 8 = 17; pre-

caudal vertebrae 10; caudal vertebrae 15 or 16; branchiostegal rays 7;

lower gill rakers 7 to 11; upper gill rakers 17 to 23; lower and upper

gill filaments increasing in number with age to a maximum, respectively,

in excess of 35 and 80; base, especially interior surface of pectoral

fin black; spinous and soft dorsal fins dusky due to melanophores on

interradial membrane; body dusky dorsally, lightly metallic ventrally;

four to nine lateral body bars, which fade with age; peritoneum flesh-



The type-species of this genus is bicolor GOnther (1860: 942),

based on two juvenile specimens. The species marginatus Bleeker (1862:

138-140), (on which Bleeker based the description of Hemicaranx), based

on one adult, is recognized as a junior synonym. The conspecificity of

these taxa is confirmed by the large number of "transitional characters"

shared between juveniles and adults. Bleeker was correct in recognizing

this form as a genus distinct from Caranx.

Carangops Gill was infrequently used in the literature by Gill,

Poey, and Goode from 1862 to 1879. It was recognized as a synonym of

Hemicaranx by Jordan and Evermann (1896: 912), although they questioned

which name was published first in 1862. However, since Carangops has

not been used in the primary zoological literature for more than 50

years it may be regarded as a nomen oblitum.

Hemicaranx has occasionally been incorrectly synonymized with

Alepes Swainson by Fowler.* However, Hemicaranx is geographically

restricted to the Atlantic and Eastern Pacific Oceans and by definition

(diagnosis and description) it is distinct from all other carangid

genera. The description of Alepes (Swainson, 1839: 248) was based on

a drawing (Russell, 1803: plate 155) of a specimen from India. As

noted by Ginsburg (1952: 97-98), however, the type species of Alepes,

A. melanoptera Swainson, is "...unidentifiable at the present time,

or else it is generally different" from Hemicaranx amblyrhynchus. Indeed,

examination of the drawing of the specimen (commonly named the "wori

parah") in Russell upon which Swainson based his generic and type

species description reveals a number of characters that support the

generic distinction hypothesized by Ginsburg. The most trenchant

differences are the number of pectoral rays (17 versus 18 to 23 in

*see species accounts annotated synonymies

Hemicaranx) (Table 17) and, more significantly, the number of principal

caudal rays (7 + 8 = 15, versus 9 + 8 = 17, the number diagnostic for

the Carangidae). Additionally, the illustration of A. melanoptera

lacks the dark pigment at the pectoral base characteristic of Hemicaranx.

It is obvious that Russell's drawing is incorrect, based on the

unusual number of principal caudal rays. Since the number is not char-

acteristic of any known carangid, the genus Alepes and the type species

A. melanoptera are regarded as nomina dubia.

The combination of characters figured for A. melanoptera were

obviously based on a carangid fish. Based on the anteriorly-arched

lateral line, scutes only on the straight posterior section, fine teeth,

body outline, and pigment in the spinous dorsal interradial membrane,

one would suppose it to be Atule malam. However, the number of first

dorsal spines (seven), number of pectoral rays, absence of pectoral-

base blotch, number of curved lateral line scales (54), and number of

scutes in the straight lateral line (44) distinguish melanoptera from

malam (Tables 20 and 21).

The utility of the inclusion by Fowler of species of both Hemicaranx

and Atule in Alepes is that it gives a historical basis for considering

the two genera as possible close relatives. This is explored below.


Since the original generic description by Bleeker, the relationships

of Hemicaranx were infrequently commented on only by North American

workers. Of the nominal genera of Carangidae from American waters

(Table 1), those most commonly cited as "close relatives" of Hemicaranx

are Caranx, Chloroscombrus, and Uraspis. Discussing Hemicaranx Ginsburg


This genus is near Caranx, differing in having a less
extensive dentition, a deeper body, and narrower maxil-
lary. It is also close to Chloroscombrus and Uraspis ....

On the basis of "characters, form and general appearance" Ginsburg (1952:

101) said Chloroscombrus is"nearest Hemicaranx," differing with fewer

and reduced scutes, more extensive dentition, and a more convex ventral

outline. Ginsburg (p. 101) also stated that of the Gulf of Mexico

carangids Uraspis is "nearest" H. amblyrhynchus, from which it differs

on the basis of forward-pointed scutes, white areas in the mouth,

longer ventral fin, lower spinous dorsal, wider maxillary, longer

lateral line arch, biseriate jaw teeth, fewer gill rakers, scutes,

and anal rays, and more pectoral rays.

Although Berry (1959: 526) did not comment on the relatives of

Caranx, he incorporated the following characters in differentiating

the two genera:

character Caranx Hemicaranx

1. maxillary width/pupil diameter greater less

2. jaw teeth different size all equal size

3. vomerine teeth present absent

4. caudal peduncle keels present absent

Indirect comments on the relationships of Hemicaranx to genera

outside its geographic range were provided by Fowler who repeatedly

included species of both Hemicaranx and Atule in the genus Alepes, the

last herein considered to be a nomen dubium. Nichols (1942b:229)

observed that Atule malam "is an approach to the carangin [sic] genus

Hemicaranx" and alluded to the hypothesis that Hemicaranx and Chloroscom-

brus may each represent an Eastern Pacific-Atlantic evolutionary lineage

from Atule. Presumably, Fowler and Nichols drew their conclusions from

overall external morphological similarity; neither, however, provided

a basis for their remarks.

Evaluation of the literature comments on Hemicaranx relationships

and/or resemblances reveals that they are often based on either (1) a

few characters that may or may not be documented as to differential

value in assigning primitivenesss" to a taxon, or (2) many characters

that yield an overall picture of morphology. In the absence of docu-

mented primitive characters, a wiser course in hypothesizing evolution-

ary trends and relationships is to assess relationships on degree of

resemblance or affinity. Sokal and Sneath (1963: 48) discuss this

approach; Gilbert (1964: 116), who followed it, stated:

The closest relatives of Luxilus are those species of
Notropis that share with it the largest number of similar
or identical morphological characters. To base a relation-
ship on only one or two shared features can be misleading.

One means of implementing such a conceptual approach is to quantify

character states according to the techniques of numerical taxonomy.

Therefore, in assessing the relationships of Hemicaranx, phenetic

resemblance, based on comparison of many random characters, is inter-

preted to reveal relationship.

A lack of critical analysis and definition of carangid species

and genera based on external morphology (Mansueti, 1963: 57; Berry,

1968: 164) restricts the number of such characters that may be coded

in such an analysis. However, the osteological catalog of Indo-Pacific

genera provided by Suzuki (1962) contains enough characters for gen-

eration of coefficients of association and subsequent description of

phenetic resemblances. Coefficients of association between Hemicaranx

and Uraspis, Caranx, and Atule are presented in Table 22. Chloroscombrus,

the other genus historically referred to as a relative of Hemicaranx,

is compared in Table 23.

Clearly, Hemicaranx most closely resembles Atule when random non-

weighted characters are compared. The two genera share a matching co-

efficient of .89, which is higher than any other coefficient between

Hemicaranx and other carangids (Table 22). Thus the hypothesis that

Hemicaranx and Atule are closely related -- whether based on subjective

impressions of overall morphology or consideration of a relatively few

characters -- is confirmed. Of the primitive characters listed in Table

6, Hemicaranx and Atule share all but one, indicating a close relationship

if only "weighted" characters are used, too.

Of the three other genera said to be close to Hemicaranx, Uraspis

ranks highest with a coefficient of .80, followed by Chloroscombrus at

.77, and Caranx at .63. It would appear that these genera are not as

closely related to Hemicaranx as originally hypothesized, especially

in light of the affinity of Hemicaranx to a group of genera at or above

the .80 level. These include Longirostrum, Selar, and Selaroides (.82-

.83), and, at .80 Trachurus, and Gnathanodon (Table 22). With the ex-

ception of Gnathanodon, all of these genera were said by Suzuki (1962:

133) to be closely related to Atule. Generation of association coeffic-

ient matrices for the above genera yields a dendrogram that illustrates

the results of cluster analysis for the closest relatives of

Hemicaranx (Figure 3). From this analysis it is concluded that Hemicaranx,

Atule, and Selar are intimately related, and as a group are most closely

related to the genus-pair of Decapterus and Trachurus. It is hypothe-

sized that Hemicaranx, based on its distribution and the close resemblance

of its species, is relatively young and is derived from Atule, or else

they have evolved from a common Indo-Pacific ancestor. Atule is further



w X X:Xcn
C3UJ 0 U0 X I
I- Q WJ In = 0
- u/ O Q ^ i
o Cn o cC W (n

Qr < Z < -1 < < < -1
z3 (r o z L W O|- i 0. 0 Ir
< -C 3 0 W : L I cc O
Cj .j C.9 Cn L n C3 U .

Figure 3. Phenetic dendrogram of genera
of Carangidae sharing a matching
coefficient (Ssm) of at least
.75 with Hemicaranx. (All Ssm's
given in Table 22)

90 o-

70 1-


hypothesized to be older, on the basis of greater morphological diver-

gence of its species.

Hemicaranx amblyrhynchus (Cuvier)

Figures 8, 19

Caranx amblyrhynchus. -- Cuvier, in Cuvier and Valenciennes, 1833:

100, pl. 248 (original description; type locality: Brazil;

type material: MNHNA5843,two syntypes, 137 and 139 mm SL). --

Gunther, 1860: 441-442 (C. falcatus a synonym; distinguished

from C. bicolor; description; range). -- Poey, 1861: 344

(comparison of amblyrhynchus from Brazil with C. heteropygus

from Cuba). -- Bleeker, 1862: 140 (comparison with Hemicaranx

marginatus). -- Bleeker, 1863: 82 (comparison with Hemicaranx

marginatus). -- Poey, 1866: 328 (distinguished from C.

heteropygus). -- Poey, 1867: 164 (distinguished from C.

heteropygus). Poey, 1875: 152 (distinguished from Carangops

heteropygus). -- Goode and Bean, 1882: 237 (Gulf of Mexico). --

Jordan and Gilbert, 1882: 308 (relationship to C. atrimanus;

comparison; C. falcatus synonymized). -- Jordan and Gilbert,

1883: 194, 197 (key; synonymy; C. secundus a synonym; Cape

Hatteras to Brazil). -- Jordan, 1884a: 34 (synonymy; Pensacola,

Florida). -- Jordan, 1884b: 284 (relationship of C. leucurus

with amblyrhynchus).

Carangops amblyrhynchus. -- Gill, 1862: 435 (compared with C.

falcatus; Brazil). -- Poey, 1868: 366-367 (C. heteropygus


Hemicaranx amblyrhynchus. -- Jordan and Evermann, 1896: 912-913

(key; synonymy; description; comparison with Hemicaranx

atrimanus; Cape Hatteras to Brazil; West Indies). -- Jordan

and Evermann, 1898: 2844 (Hemicaranx falcatus recognized as

distinct from amblyrhynchus).-- Jordan and Evermann, 1900:

plate CXL1 (Figure 386).-- Nichols, 1922: 59 (description

'of juveniles; comparison with H. marginatus from West Africa;

ecological association with medusae; Miami Beach, Florida).--

Meek and Hildebrand, 1925: 339 (key; synonymy; description;

Fox Bay, Colon, Panama).-- Jordan Evermann and Clark, 1930:

271 (synonymy; range).-- Nichols, 1937: 5-6 (juvenile growth

patterns; Caranx falcatus and Hemicaranx rhomboides synonyms;

comparison with H. marginatus; South Carolina).-- Howell Y

Rivero, 1938: 56 (Caranx heteropygus identified as H. amblyrhyn-

chus).-- Hildebrand, 1941: 226 (Beaufort Inlet to Cape Lookout,

North Carolina).-- Nichols and Murphy, 1944: 242 (H. rhomboides

synonymized; comparison with H. leucurus).-- Baughman, 1947:

280 (Aransas Bay, Texas).-- Irvine, 1947: 139 (tentatively

synonymizes H. bicolor, West Africa).-- Breder, 1948: 131, 135

(key;association of young with jellyfish in lower Mississippi

R.; Cape Hatteras to Brazil).-- Baughman, 1950: 245 (small

specimens under Aurelia jellyfish; color notes; Texas).--

Ginsburg, 1952: 98-99, 101 (synonymy; description; close

relatives are Uraspis heidi and Chloroscombrus chrysurus;

northern Gulf of Mexico).-- Matthews and Shoemaker, 1952: 270

(small individuals observed and collected in and with jelly-

fish medusae and the ctenophore Beroe; Mississippi Sound,

Biloxi).-- Hildebrand, 1954: 301, 328 (young often found under

bell of cabbagehead jellyfish, Stomolophus meleagris; trawl

and trynet collections at Greens Bayou, and Pass Cavallo to

Colorado River, Texas; "not uncommon in northern Gulf").--

Reid, 1955: 440 (in trawl; salinity 17.6-24.3 ppt; East Bay,

Texas).-- Boeseman, 1956: 193 (description; Cape Hatteras to

Brazil; first Surinam record, from near lightship "Surinam

River").-- Joseph and Yerger, 1956: 133, 156 (Alligator Harbor,

Florida; latitudinal range).-- Reid, 1956: 316 (in trawl;

not in seine or trammel net; East Bay, Texas).-- Springer

and Bullis, 1956: 75 (Oregon collections: 30017'N, 88029'W;

30016.3'N, 88029'W; 3015'N, 88025'W).-- Reid, 1957: 207

(East Bay, Texas).-- Briggs, 1958: 277 (pelagic; Florida;

range).-- Hoese, 1958: 334 (ecological association with jelly-

fish; Texas).-- Berry, 1959: 525 (questions relationship to

Poey's cotypes of Caranx secundus).-- Smith and Bailey, 1961:

359 (predorsal formula: 0-0-0-1-).-- Mansueti, 1963: 56

(young specimens taken with jellyfish, Chrysaora quinquecirrha,

Gulf of Mexico, near Biloxi, Mississippi, Stomolophus meleagris,

Texas, Aurelia aurita, Texas, Mastigias scintillae, Sao Paulo,

Brazil, and unidentified species, Miami Beach, Florida, and

Texas).-- Bullis and Thompson, 1965: 41 (Oregon collections:

28050'N, 87058'W; 29040'N, 93023'W).-- Copeland, 1965: 17-18

(ecological association with cabbagehead jellyfish, Stomolophus

meleagris; occasional fall and winter emigration through

Aransas Pass, Texas).-- Parker, 1965: 212 (salinity 10-35 ppt;

uncommon Galveston Bay system, Texas).-- Roithmayr, 1965: 21

(in industrial bottomfish trawl catches; area: 280-300N, 87030'-

90030'W).-- Cervigon, 1966: (Venezuela).-- Berry, 1968: 148

(number of vertebrae).-- BShlke and Chaplin, 1968: 322 (not

collected, but expected in Bahamas).-- Gines and Cervigon,

1968: 33 (6048'N, 57038'W; 6040'N, 57021'W; 6034'N, 57011'W;

618'N, 55055'W).-- Randall, 1968: 102 (not collected, but

"common" in the Caribbean).-- Phillips et al., 1969: 703

(ecologically associated with sea nettles and ctenophores,

Beroe; Mississippi Sound).-- Bailey et al., 1970: 40 (U.S.

Atlantic; common name "Bluntnose jack").-- Roessler, 1970:

863, 884 (Buttonwood Canal, Everglades Park, Florida).

Alepes amblyrhynchus.-- Fowler, 1905: 71-72 (description; Rio de

Janeiro, Brazil).-- Fowler, 1936: 690, fig 310 (synonymy of

Hemicaranx marginatus and Caranx bicolor with amblyrhynchus;

description; tropical Atlantic).-- Fowler, 1941: 153 (Brazil).

-- Fowler, 1945: 189, 375 (synonymy; South Carolina, Texas).

Caranx falcatus.-- Holbrook, 1855: 92-94 (original description; type

locality Charleston, South Carolina; one specimen; comparison

with C. amblyrhynchus).-- Holbrook, 1860: 94 (description;

Charleston, South Carolina).-- Poey, 1867: 165 (distinguished

from C. heteropygus).-- Poey, 1875: 152 (distinguished from

Carangops falcatus).-- Nichols, 1937: 6 (synonymized with

Hemicaranx amblyrhynchus).

Carangus falcatus.-- Gill, 1861: 36 (eastern coast of North America).

Carangops falcatus.-- Gill, 1862a: 238 (variation in dentition; in-

distinguishable from C. heteropygus).-- Gill, 1862b: 431, 435

(key; type of Carangops Gill; compared with C. amblyrhynchus

from Brazil; Charleston, South Carolina).-- Goode, 1879: 112

(East coast of Florida).

Hemicaranx falcatus.-- Jordan and Evermann, 1898: 912 (description;

distinguished from amblyrhynchus; Charleston, South Carolina).

-- Jordan, Evermann, and Clark, 1930: 271 (Charleston, South


Caranx heteropygus.-- Poey, 1861: 344, 373 (original description;

Havana market; one specimen, MCZ 17254; compared with C.

amblyrhynchus).-- Poey, 1866: 328 (distinguished from C.

amblyrhynchus; Cuba).-- Poey, 1867: 164-165 (distinguished

from C. amblyrhynchus and C. falcatus; Cuba).-- Howell y

Rivero, 1938: 56 (type specimen, MCZ 17254, identified as

Hemicaranx amblyrhynchus).

Carangops heteropygus.-- Gill, 1862a: 238 (indistinguishable from

C. falcatus).-- Poey, 1868: 366-367 (synonymized with C.

amblyrhynchus; Cuba).-- Poey, 1875: 151-152 (distinguished

from C. amblyrhynchus and C. falcatus; Cuba).

Hemicaranx rhomboides.-- Meek and Hildebrand, 1925: 343, pl. 25, fig. 2

(original description; type locality, Fox Bay, Colon, Panama;

type material: USNM 81758, 2 specimens, SL 55 and 75 mm;

compared with H. leucurus and H. secundus).-- Jordan, Evermann,

and Clark, 1930: 271 (Atlantic coast of Isthmus of Panama).--

Nichols, 1937: 6 (synonymized with H. amblyrhynchus).-- Nichols

and Murphy, 1944: 242 (comparison with H. leucurus; synonymized

with amblyrhynchus).-- Gunter, 1945: 57 (juveniles; one in

trawl, five in seine, summer; Aransas Bay, Texas).-- Grey, 1947:

154, 201 (one paratype located in Field Museum of Natural


Material Examined

United States

North Carolina.-- USNM 111787 (1, 57.3), Carteret Co., Cape Lookout

Bight, J. S. Gutsell, 2 Sept. 1927; USNM 164487 (1, 19.8), Carteret

Co., Beaufort, J..S. Gutsell, 2 March 1933; USNM 112746 (1, 36.8),

Carteret Co., Beaufort, Sea Buoy, J. S. Gutsell, 17 Oct. 1931;

USNM 112745 (1, 45.5), Carteret Co., Beaufort, W. Bell Buoy, 13

Sept. 1914; USNM 112747 (1, 42.7), Carteret Co., Fort Macon, outer

beach, otter trawl, 31 July 1916.

South Carolina.-- USNM 155275 (1, 102.3), Charleston Co., off Bull

Bay, 18 Oct. 1937; AMNH 13648 (2, 93-98), Charleston Co., Bull Bay,

trawl, E. M. Burton, 26 Aug. 1936; CM 36.165.8 (6, 82-99), Charles-

ton Co., Bull Bay, trawl, E. M. Burton, 26 Aug. 1936; AMNH 13647

(1, 73.0), Charleston Co,, n. end of Cape Island, E. M. Burton,

12 Aug. 1936; CM 36.164.6 (3, 61.5-67), Charleston Co., n. end of

Cape Island, E. M. Burton, 12 Aug. 1936; CM 35.322.2 (1, 58.5),

Charleston Co., Morris Island, E. M. Burton, 22 Sept. 1935; CM 31.

190.11 (1, 75.5), Charleston Co., Stone Inlet, trawl, John T.

Nichols, 12 Aug. 1931; CM 38.201.1 (2, 87-94.5), Charleston Co.,

north jetty, Charleston, trawl, E. M. Burton, 21 Aug. 1938; CM 38.

201.1 (2, 88.5-93), Charleston Co., Charleston north jetty, E. M.

Burton, 21 Aug. 1938; CM 34.175 (2, 154-157), Charleston Co.,

Charleston Harbor, E. L. Passailague, 9 July 1934; USNM 5990 (2,


Georgia.-- USNM 119232 (1, 81), Glynn Co., St. Simons I.; TABL

105333 (1, 112), Glynn Co., Commercial trawling area off Brunswick,

Ga. ca. 31007'N, 81010'W, Lewis Crab Co. shrimp trawl, 20 Oct.

1955; TABL 105334 (1, 98), Glynn Co., off Jekyll Island, ca. 31004'N,

81023'W, Jane Briggs, J. E. Karr shrimp trawl, 9 July 1959; TABL

105332 (1, 158), Glynn Co., Jekyll Island, E. (ocean) side, ca.

31004'N, 81024'W, Lewis Crab Co. shrimp trawl, 10 June 1957; TABL

105331 (1, 68), Doboy Sound, ca. 31022'N, 81015'W, Doc. Jones -

Ga. Game and Fish Comm., 2 Aug. 1957.

Florida.-- FSBC 1015 (2, 80.4-86.3), Duval Co., Jetty at Atlantic

Beach, 29 Nov. 1958; TABL 105330 (1, 123), Florida Atlantic, 30 31'N,

81022'W, 7 fms, 40' 2-seam trawl Silver Bay, 5 Oct. 1961; AMNH

8092 (8, 21.5-57.5), Dade Co., Miami Beach, L. L. Mowbray, 27 July

1921; USNM 39873 (1, 88.6), Monroe Co., off Cape Sable, Moser; SU

36285 (1, 61.4), Lee Co., Sanibel, M. Storey, Spring, 1933; FSBC

2071 (1, 168), Pinellas Co., Indian Rocks Pier, Indian Rocks Beach,

4 June 1961; FSU 1377 (4, 55.8-65), Franklin Co., Alligator Harbor,

16-22 Sept. 1952; FSU 5698 (10, 57.2-82.7), Franklin Co., Mud Cove

off Alligator Peninsula, W. Menzel, 3 Oct. 1959; SUCAS 767 (1, 221),

Escambia Co., Pensacola.

Mississippi.-- USNM 155274 (1, 79.9), Harrison Co., off Gulfport,

23 Sept. 1939.

Louisiana.-- USNM 100718 (1, 149); Jefferson Parrish, just off

front (south) beach Grande Isle; GCRL 265 (1, 57.5), Jefferson

Parrish, S. Grand Isle 3.5 fms., Dawson trawl field 582, 22

Nov. 1959; AMNH 14217 (14, 23-42), Cameron Parrish, Calcasieu R.,

Townsend, 12 Aug. 1928; GCRL 195 (1, 40), Gulf of Mexico, 29040'N,

9323'W, 5-5.4 fms., M/V Oregon Sta. 2875, Field No. 574, 7 Aug.


Texas.-- USNM 120055 (1, 102), Galveston Co., Galveston, J. L. Baugh-

man; TABL 105329 (4, 61-80), Brazoria Co., Freeport, Mar. Lab.

Mus., Texas Game, Fish and Oyster Comm., Apr.-Sept. 1947; BMNH 1948.8.6.

478.479 (2, 64-75.8), Aransas Co., Aransas Bay, Baughman; USNM

144014 (1, 62.6), Aransas Co., Aransas Pass, Harbor Island, J. C.

Pearson, 6 Dec. 1926; USNM 119805 (1, 96.4), Nueces Co., Corpus

Christi, J. C. Pearson; USNM 144072 (2, 118-133), Nueces Co.,

Corpus Christi Bay, J. C. Pearson, Nov.-Dec. 1926; ANSP 70602 (1,

121), Nueces Co., Corpus Christi, 1931.

Gulf of Mexico.-- GCRL 1251 (1, 196), Eastern Gulf of Mexico -

2-7 fms., M/V Tony, 2 Aug. 1962; UF 3931 (1, 32.9), S. Mobile,

Ala., 28 50'N, 87058'W, Oregon 1593, D. K. Caldwell, 25-26 July

1956; TU 4355 (1, 131), S. Horn Is., Miss., 300163'N, 88029'W,

2.8 fms., Oregon 627 36-65 mid-water trawl, 27 Aug. 1952; TU

4360 (1, 180), S. Ship I., Miss., 30017'N, 88051.6'W, 2.7 fms.,

Oregon 622 35-65 mid-water trawl, 23 Aug. 1952; TABL 102645

(2, 38.1-53.8), S. Cameron Parrish, 29031'N, 92045'W, Silver Bay

No. 2869, 6 Aug. 1960.


MCZ 17254 (1, 268), holotypee of Caranx heteropygus Poey).


TABL 101355 (1, 151), east, parallel to beach off Caratasca Lagoon,

past Rio Cruta, 15019'N, 83026'W, 5 fms., UN 6703 Shady Lady 60'

trawls double rig, G. C. Miller, 10 April 1967; TABL 101400 (1, 70),

east of Rio Cruta, 15018"N, 83022'W, 5 fms., UN 6703 Shady Lady

try net, G. C. Miller, 10 April 1967; TABL 102757 (4, 116-166),

Commercial trawling area off Caratasca Lagoon east to past Rio

Cruta, 15019'N, 83026'W, 5 fms., UN 6703 Shady Lady 60' trawls, dou-

ble rig, G. C. Miller, 10 April 1967; TABL 101356 (7, 128-167),
west past Rio Cruta River to off Caratasca Lagoon, 15 21'N, 83034'W,

5-6-1/2 fms., UN 6703 60' trawls double rig, G. C. Miller, 10

April 1967; TABL 105388 (5, 127-167), Rio Cruta-Caratasca, 15026'N,

83041'W, 5-6 fms., UN 6703 60' trawls, double rig and try net Shady

Lady, G. C. Miller, 11 April 1967.

Costa Rica

LACM 30727 (1, 128), Cahuita Bay, W. Bussing.


USNM 80007,(1, 175), Colon; USNM 81758 (1, 58.5), Colon, Fox Bay,

22 Jan. 1912.


TABL 102891.(4, 172-180), 11033'N, 71031'W, 15 fms., 104 Oregon

40' shrimp trawl, 6 Oct. 1965; TABL 105035 (1, 215), off Puerto

la Cruz, 10O0l'N, 64048'W, 19 fms., 40' flat trawl, 19 Oct. 1963.


BMNH 1931.12.5.167-8.(2, 63.1-120), Gulf of Paria, Talton, Rodney;

ANSP 86221 (3, 169-184), Port of Spain, Barber Asphalt Co., 1930;

BMNH 1932. (3, 102-199), Gulf of Paria, Guppy; USNM 178437.

(1, 38.2).

British Guiana

TABL 104329 (1, 168), 08045'N, 59015'W, Trawl Calamar, 18 June 1967;

TABL 104383 (1, 155), 07015'N, 58015'W, 67-6 Calamar Trawl, 20

June, 1967; BMNH 1961.9.1.25-26 (2, 138-174), R. M. McConnell.


USNM 220144,(2, 196-205), 07012'N, 56047'W, 26-28 fms., Oregon Sta.

2263-80 ft. balloon trawl, 1 Sept. 1958; TABL 105335 (1, 57)

06018'N, 55030'W, 7 fms., 40' flat trawl R/V Oregon Sta. 2280,

4 Sept. 1958; TABL 104788 (2, 197-207), 06015'N, 54045'W, 67-11

Calamar Trawl, 3 Nov. 1967; TABL Uncat. (2, 154-160), NW coast, 12.5-

14 fms., 68-11 Calamar, Dec. 1968; TABL Uncat. (2, 176-177), NW

coast, Calamar Cruise 68-11, Sta. 591, 12 Oct. 1968; RMNH Uncat.

(1, 65 ), 38 ft., Coquette Stat. 7 trawl, 29 June 1966; RMNH 18148

(3,, 53.2-63.7), Surinam R. near Plantation Resolutie; RMNH Uncat.
(1, 121), 7.7 mi. E. lightship Surinam River, 40-50 ft., 29 June

1966; RMNH 21478 (4, 51.5-83), off lightship Surinam River; RMNH

24771 (2, 69.9-74.7).


Para.-- CAS-SU 22113 (1, 152), E. C. Starks.

Ceara.-- CAS-SU 51879 (1, 126), Port of Fortaleza (Mucuripe), 23

Feb. 1945; CAS-SU 51824 (4, 26.7-43.5), Fortaleza (Mucuripe),

March 1945.

Pernambuco.-- CAS-SU 67017 (1, 210), Recife, N. Berla, 8 Aug.

1944; CAS-SU 67026 (1, 211), Recife.

Bahia.-- BMNH 1844.5.14.63 (1, 147), Salvador, Parzudakis Colln.;

BMNH 1862.1.30-20 (1, 171), Salvador; CAS-SU 66984 (1, 123),

Maranhao and Bahia, Salvador, 1944 or 1945; CAS-SU 67023 (1, 178),

Salvador; CAS-SU 67025 (1, 255), Salvador.

Sao Paulo.-- CAS-SU 66998 (1, 113), Ponto do Praia, Santos; CAS-SU

34813 (1, 123), Santos, A. W. Here, 3 June 1934; CAS-SU 66997

(3, 95-121), Ponto Do Praia, Santos, P. Carvallo and Moraes, 14
March 1944; CAS-SU 66992 (2, 111-119), Santos; CAS-SU 66986 (1,

94), Ponto do Praia, Santos, V. Carvalka and Moraes, 23 May 1943.

Rio de Janeiro.-- BMNH 1923.7.30.145-6 (2, 133-139), Rio de Janeiro

(fish market), Ternetz; ANSP 11259 (1, 148), Rio de Janeiro.

Santa Catharina.-- CAS-SU 67019 (1, 182), Florianopolis, Praia de

Carrosvicira, 19 Oct. 1943.


Hemicaranx amblyrhynchus is distinguished from other members of

its genus by the following combination of characters: upper caudal

fin lobe extremely long, nearly 50% SL in largest adults; upper caudal

fin lobe up to 30% larger than lower lobe length; ventral body outline

slightly convex in advance of soft anal fin origin, but not as broadly

rounded as dorsal outline; anal pterygiophores closely spaced.

It is further distinguished from the closely related H. bicolor

by the following characters: more dorsal rays (24 to 30, usually 27

or 28) (1 + Sx = 27.62 + 0.40); more anal rays (21 to 25, usually 23

to 25) ( + Sx = 23.74 + 0.35); a flattened ascending process (versus

indented) and concave articular process of the premaxillary (versus

straight); basihyal of intermediate width (versus narrow).

H. amblyrhynchus is further distinguished from both H. leucurus

and H. zelotes by the following characters: seven (versus eight) dorsal

spines; 16 (versus 15) caudal vertebrae; fewer scales in the curved

part of lateral line (25 to 44, usually 34 to 39) (X + Sx = 37.86 + 1.18)

(versus 29 to 54); anterior caudal peduncle scute proportional width 3.9

to 4.8% SL ( = 4.4) (versus 1.7 to 3.6); ratio of straight portion of

lateral line to curved portion of lateral line 2.3 to 3.0 (X = 2.6)

(versus 1.9 to 2.3); antero-dorsal edge of ethmoid concave (versus

convex); anterior end of pterygoid bone indented (versus concave);

upper hypohyal window circular (versus oval); anal pterygiophore points

intermediately developed (versus prominent or reduced).

A comparison of H. amblyrhynchus and the other members of the genus

is presented in Table 25.


Counts and proportional measurements are listed in Tables 7, 8,

15-19, and graphically presented in Figures 5-7, 9, 11, 14, 22. The

generic description and species diagnosis are supplemented by the

following: length of straight lateral line 50 to 60% SL; length of

curved lateral line 20% SL; body width 13% SL; pectoral fin 30% SL;

pelvic fin 33% pectoral; head depth 30% SL; interorbital width 9% SL;

maxillary depth 2% SL; pectoral fin rays 18-21, usually 19 or 20;

scutes in straight part of lateral line 34 to 51 (X + Sx = 44.87 + 0.29);

distal tips of the upper superior principal caudal rays lightly blacken-

ed in adults.

Osteological characters are completely described in the account

of the osteology of the genus. Skeletal characteristics that disting-

uish H. amblyrhynchus from other members of the genus are: anterior

edge of dorsal surface of ethmoid concave; posterior tip of upper arm

of dentary rounded; posterior edge of postmaxillary process broadly

concave; a mid-dorsal expansion of the symplectic; pterygoid character-

ized by indentations above and below anterior point; upper hypohyal

foramen circular; ceratohyal window ventrally expanded; anal pterygial

points moderate; distal ends of anal pterygiophores closely opposed.


The original description of Caranx amblyrhynchus Cuvier (commonly

named the Bluntnose jack [Bailey et al., 1970: 40]) was based on two

syntypes from Brazil. Both specimens agree with the original descrip-

tion. In accord with Bailey (1951), authorship of the specific name

is recognized.

Caranx heteropygus Poey was based on a single specimen from Cuba.

The taxonomic status of heteropygus vacillated, even in the works of

Poey, who distinguished it from amblyrhynchus strictly on the basis

of upper caudal fin lobe length. This character is within the normal

range of variation of amblyrhynchus, however, and all other morpho-

metric characters agree with the data obtained for that species.

Howell Y Rivero (1938: 56) correctly identified Poey's type specimen

as amblyrhynchus.

The original description of Hemicaranx rhomboides Meek and Hilde-

brand was based on two juvenile specimens from Caribbean Panama. One

of these (the holotype) was found in the USNM type collection and it

agrees in every way with amblyrhynchus. The paratype is deposited in

the Field Museum of Natural History (Grey, 1947: 154, 201). Meek and

Hildebrand (1925: 342) examined only juvenile specimens, which differ

in many ways from the adult specimens upon which their account of

amblyrhynchus was based. Nichols and Murphy (1944: 242) correctly synonym-

ized rhomboides with amblyrhynchus.

The original description of Caranx falcatus was based on a single

adult specimen collected twenty miles offshore from Charleston, South

Carolina. The etymology of the name reflects the major distinction

Holbrook incorporated in a comparison of his specimen with the figure of

a syntype of Cuvier: i.e., the relatively longer length of the upper

caudal fin lobe (most pronounced in specimens from more temperate

localities). This variation is characteristic of amblyrhynchus, and

in this and all other details the type specimen of falcatus agrees with

amblyrhynchus. The specimen was accidentally destroyed (Holbrook, 1860:

96). Falcatus was correctly synonymized with amblyrhynchus by Nichols

(1937: 6).

An additional species originally described as Caranx secundus by

Poey (1860: 223) was doubtfully synonymized with amblyrhynchus by

Jordan and Gilbert (1883: 197). This species is now regarded as a

number of the genus Uraspis (Berry, 1963: 584). In the same paper,

Berry noted that Caranx fasciatus, a nomen dubium, has at times been

regarded as a synonym of secundus and as a species of Hemicaranx.


Hemicaranx amblyrhynchus has been referred to by Briggs (1958: 277)

as a "pelagic" species, inhabiting

the surface layers of water -- depths of less than
200 meters -- in the offshore regions usually beyond
the limits of the continental shelf.

The bulk of the specimens of amblyrhynchus examined in this study, how-

ever, were collected in relatively shallow waters adjacent to continental

land masses, thus indicating a life cycle at least partially spent in

waters less than 200 meters deep. Despite such factors as (1) lack of

offshore collecting effort and (2) escape from collecting gear that may

bias conclusions about habitat, the almost exclusive collection of

Hemicaranx amblyrhynchus in shallow waters over the continental shelf

identifies it as a "shore" species (see Briggs, 1958: 277). Copeland

(1965), in a discussion of animal emigration through Aransas Pass, Texas,

alluded to the estuarine dependence of this species. In a year-round

tide trap sampling of emigrants through Aransas Pass, Copeland (1965:

Table 1) found amblyrhynchus to occur occasionally in collections made

from October through February. Although Copeland did not mention the

size of specimens collected, they are presumably juveniles, based on

the collecting gear and a consideration of other collections from Texas.

Gunter (1945: 57), for example, reported Bluntnose jack of 32 to 82 mm

from Aransas Bay during June, July, and August.

Whether amblyrhynchus spawns in estuarine habitat is not certain,

but it appears that this species is one of many that utilizes the estuar-

ine environment as a nursery ground for juveniles. From an estuarine

nursery it moves offshore to complete part or all of the life cycle. As

listed in the Material Examined section, juveniles of amblyrhynchus have

been collected at a number of other estuarine localities, whereas adults

are always collected in offshore (full-strength sea) waters.

The ecological association of juvenile H. amblyrhynchus with jelly-

fishes has been reported from a number of localities throughout its

range (summarized by Mansueti, 1963: 56-57), and is by no means unusual

for the Carangidae. Indeed, Mansueti (1963: Table 5) listed published

records of commensal and parasitic association between jellyfishes and

the young of 23 species of Carangidae in the world. For amblyrhynchus

Mansueti listed the following symbionts: Chrysaora quinquecirrha,

Stomoloplus meleagris, Aurelia aurita, and Mastigias scintillae.

Interestingly, amblyrhynchus was collected by Copeland (1965: 18) only

in association with the cabbagehead jellyfish, S. meleagris, during

fall emigration from Aransas Bay.

In his review article, Mansueti stated that such fish-jellyfish

association may (1) protect the fish from predators and (2) provide

food in the form of crustaceans and other invertebrates found on or

with the jellyfish, as well as the jellyfish itself. Immunity of

carangids to nematocyst toxin was suggested by Mansueti (1963: 54).

The function of pelagic symbiotic hosts as dispersal mechanisms is

discussed below in the Relationships section, as a factor in possible

interspecific gene flow in Hemicaranx.


Hemicaranx amblyrhynchus ranges from Beaufort, North Carolina,

south along the continental margin of the Atlantic coast of North

and South America to the vicinity of Florianopolis, Brazil (ca. 28 S)

(Figure 4). Although it may be common in occasional samples, H.

amblyrhynchus is most often present in institutional collections in

small numbers. Although this paucity of specimens may reflect

selectivity of sampling techniques or a failure to sample preferred

environmental habitat, it appears that this species is truly uncommon

in the natural environment, since even the more easily collected

juvenile stages are never taken in abundance. Tide traps collections

(Copeland, 1965; Roessler, 1970), beach seining, and trawling have all

failed to capture large numbers of amblyrhynchus.

Along continental shores H. amblyrhynchus is differentially

abundant, no doubt partly as an artifact of collecting intensity. How-

ever, this species -- although expected by various authors (Bshlke and

Chaplin, 1968: 322 Bahamas; Randall, 1968: 102 non-inshore Caribbean

saline habitats) -- has not been reported from even the most actively

collected islands of the northern and eastern Caribbean Sea. This is

interesting in view of the common occurrence of many other carangids

in such localities, including such closely related forms as Selar

(Randall, 1968: 106), with which Hemicaranx is sympatric over much

of the continental shelf. Both Gilbert (in press) and Robins (1971:

254) note that the Carangidae as a family is ubiquitously distributed

in both continental and insular waters which are characterized respect-

ively by turbid waters and notable environmental fluctuations or clear

waters and stable environmental conditions. Based on the apparently


0 0 0 0 0 0
f C 0 C a





o0 0

o -o C

- X 0 ll

L- C0
U .

a a

04 -U
o 0*

L o .
'0 04J

o 4)

so U

1 I- I- 0

N -o o


small population size of this species, plus its apparent estuarine.

dependence, it is likely that amblyrhynchus does not inhabit most of

these islands, possibly due to a lack of extensive estuarine nursery

ground complexes. The absence of amblyrhynchus from the Caribbean

Islands parallels that of the sea turtle, Lepidochelys, which Pritchard

(1969: 182-183) accounted for on the basis of a paucity of extensive

brackish water areas.

Undoubtedly, however, these medium-sized, highly mobile fish have

had access to such habitats by virtue of obvious carangid morphological

adaptations for open-water locomotion (e.g., narrow peduncle, broad caudal

lobes, compressed terete body). In addition, the symbiotic association

of juvenile amblyrhynchus with jellyfishes is a likely factor in effect-

ing its widespread dispersal. Not all free-swimming large species, of

course, are able to inhabit the entire spectrum of environments to which

they have access. The Spanish mackerel, Scomberomorus maculatus, for ex-

ample is absent from insular environments of the western Atlantic,

doubtless because, according to Robins (1971: 252):

...geographic barriers are not operating to restrict
the continental element from the islands but instead...
ecological conditions and competition from closely
related and better adapted island species provide the
barrier to colonization. What we have in these faunas
is not a picture of what could reach the area in question
but of what could survive and breed there.

It is hypothesized that H. amblyrhynchus is one species of the continent-

al fauna to which this statement applies.

The vagility of groups such as the Carangidae has been discussed

by Rosenblatt (1963) as a factor in the relative lack of endemism in

this family. As contrasted to smaller fishes adapted to discontinuous

isolated habitats, which are restricted in their mobility (and thus

gene exchange), the carangids are larger and more mobile and generally

inhabit continuous habitats over which they are free-swimming. Thus,

opportunities to promote gene flow in this group contribute to what

may be a differentially slower rate of local differentiation. This is

certainly the case for Hemicaranx amblyrhynchus.


Samples of H. amblyrhynchus from throughout the range of the species

are remarkably homogeneous as indicated by inspection and statistical

analysis of data (see in Tables 7, 8, and 15-17 in part). The small

amount of variation noted is clinal in nature: samples from more

temperate latitudes are characterized by longer caudal fin lobes (Figure

5, Table 7), higher numbers of soft rays in the second dorsal fin

(Figure 6, Table 15) and higher numbers of soft rays in the anal fin

(Figure 7, Table 16). Based on the classical observation of more body

parts in fishes from colder environments (Barlow, 1961), such a pattern

of geographic variation of meristic characters might be predicted.

Although it is tempting to conclude that nearly equatorial popu-

lations of amblyrhynchus (e. g., Guyanas-Brazil) develop fewer dorsal

and anal rays than more temperate populations (e.g., U. S. Gulf of

Mexico) in response to warmer developmental temperatures, the difference

between Honduras and equatorial localities with similar temperatures

must be accounted for. Samples from Honduras and the U. S. Gulf

(localities with different thermal regimes [Table 24]) are nearly

identical meristically, whereas samples from Honduras and the equatorial

localities are statistically significantly different (o( = .05) in

terms of dorsal and anal ray counts. Thus, if these carangid fishes

are able to disperse widely to inhibit divergence of temperate and

4 0o





,9 0

g (

. 1 I I m I



0 0 00 0 0 0 0 0 0
o '\O L ~ C C -m

(N3 3 NI)3dd
H19N31 3801 NI IVGnV2 y3ddn









o E

C) F-

o "L


o Q









0 0






^- --


O- 0










25 26 27 28 29


Figure 6. Variation in number of fin rays in the soft dorsal
fin of Hemicaranx. (Vertical line indicates x,
horizontal lines extend to either side for a distance
equal to S N.) (N=9;o(=.05 [after Eberhardt, 1968:
fig. 2])









bicolor -


Variation in number of fin rays in the soft anal fin
of Hemicaranx. (Symbols as in Figure 6; N=9;S==.05)



Figure 7.