PHYLETIC STUDIES OF

TELEOSTEAN FISHES,
- WITH A PROVISIONAL
CLASSIFICATION OF LIVING FORMS

P. HUMPHRY GREENWOOD, DONN E. ROSEN,

STANLEY H. WEITZMAN, AND
GEORGE S MYERS

VOLUME,-.,~
~~~~~~~~~~~~~~~~~~
,mlt 4...

BULLETIN
OTHE-
AMERICAN~MUSEU'M ~OF :NAmTtJAVITR
VOLUME ARTYCL 4 *.WOR1966
PHYLETIC STUDIES OF TELEOSTEAN
FISHES, WITH A PROVISIONAL
CLASSIFICATION OF
LIVING FORMS

P. HUMPHRY GREENWOOD
Assistant Keeper, Department of Zoology
British Museum (Natural History)
DONN E. ROSEN
Associate Curator, Department of Ichthyology
The American Museum of Natural History
STANLEY H. WEITZMAN
Associate Curator, Division of Fishes
United States National Museum, Smithsonian Institution
GEORGE S. MYERS
Professor of Biology and Curator of Zoological Collections
Stanford University

BULLETIN
OF THE
AMERICAN MUSEUM OF NATURAL HISTORY
VOLUME 131 : ARTICLE 4 NEW YORK 1966
BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY
Volume 131, article 4, pages 339-456, text figures 1-9,
plates 21-23, charts 1-32
Issued April 18, 1966
Reprinted August, 1968
Price: $3.00 a copy
 FOREWORD
THIS PAPER PRESENTS an attempt to assess barest beginnings have so far been accom-
the main phyletic trends in teleostean fishes, plished, even within the realm of osteology.
based primarily on study of the living forms. We doubt if more than the external anatomy
Results indicate the necessity of a major re- of 95 per cent of the species of living teleosts
grouping of teleostean orders, and this also is has been examined, and for many families
attempted. there has so far been little or no deeper study.
The work stemmed from talks and corre- Researches on the nervous, digestive, mus-
spondence between Greenwood, Rosen, and cular, and vascular systems of teleosts are
Myers at the time of and subsequent to the scattered and mostly uncoordinated, and rel-
XVIth International Congress of Zoology in latively few of them have been done with any
Washington in 1963, especially during dis- specific systematic objectives in view, de-
cussions of the then unpublished results of spite the fact that pioneer work, such as that
Greenwood's study of the osteoglossiform of Freihofer (1963) on a single nerve com-
fishes and of Rosen's work on the atherini- plex, has uncovered a wealth of information
form fishes (Greenwood, 1963; Rosen, 1964). bearing upon phyletic relationships.
Active cooperative work was not initiated
until early in 1964, when Weitzman joined ACKNOWLEDGMENTS
the group. Many of our colleagues have been exceed-
The main burden of the investigation has ingly generous in providing us with informa-
been carried by Greenwood and Rosen, tion, specimens, and even unpublished con-
closely followed by Weitzman, who is almost clusions resulting from their own research:
wholly responsible for the strictly ostariophy- Drs. James W. Atz, Reeve M. Bailey,
san lineages and for the stomiatoid work, and Frederick H. Berry, James E. Bhlke,
who has also done the figures. Myers has Charles M. Breder, Jr., Daniel M. Cohen,
acted principally as adviser and editor, and Bruce B. Collette, Warren C. Freihofer,
has contributed in large part to the compila- Robert H. Gibbs, Jr., Messrs. Richard Haed-
tion of the list of family names. Others have rich, C. M. H. Harrisson, Dr. Ernest A.
helped substantially and are thanked below Lachner, Mr. Richard Lund, Mr. Norman B.
under Acknowledgments. Marshall, Drs. Samuel B. McDowell, Giles W.
Part of the earlier findings of Rosen and Mead, Colin Patterson, Mr. Tyson J. Rob-
Greenwood were communicated orally by erts, Drs. C. Richard Robins, Alfred S.
them to A. S. Romer, during a conference in Romer, Bobb Schaeffer, C. Lavett Smith,
London early in 1964, for use by the latter in Victor G. Springer, William N. Tavolga, W.
the new edition of his "Vertebrate Paleontol- Ralph Taylor, and James W. Tyler.
ogy." That book is still in press at the time To Dr. Reeve M. Bailey we are especially
we write. grateful for having read through and criti-
Traditionally, studies such as ours have cized the entire manuscript and for having
been based on morphology, especially the provided us with an unusually large amount
skeleton, which is the only complete organ of original information on the relationships of
system available for detailed comparison with certain groups and on the validity and
fossils. However, with the variety of both orthography of many of the names used in
primitive and advanced teleosts living today, our classification.
we are most emphatically of the opinion that For editorial and technical assistance we
approaches other than morphological ones are grateful to Mmes. Carmella B. Rosen,
would be exceedingly fruitful in the investi- Marilyn Weitzman, Mary G. Hume, and
gation of teleostean interrelationships. Miss Victoria Pelton.
It must be not imagined, however, that the We feel that we, as well as all other ichthy-
full informational content of teleostean mor- ologists, owe a debt of gratitude to the late
phology has been extracted, or that it will be John Roxbrough Norman, for his technically
fully extracted for a long time. Only the unpublished (mimeographed) "Draft Synop-
341
342 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

sis of the Orders, Families and Genera of by a National Science Foundation grant
Recent Fishes and Fish-like Vertebrates," (GB-1340) to D. E. Rosen, one (GB-4685)
which we have found to be of the greatest as- to G. S. Myers, and another (GB-3906) to
sistance. R. H. Gibbs, Jr., and S. H. Weitzman
The present study was supported in part
 CONTENTS
FOREWORD . .............. ............ 341
Acknowledgments .............. ............ 341
INTRODUCTION .................. . . 345
History .............. . 345
Theory ... ................ 345
Teleostean Diversity and Age . . . . . . . . . . . . . . . . . . . . . . . . 346
NATURE OF MAJOR GROUPINGS ADOPTED . . . . . . . . . . . . . . . . . . . . 348
Major Trends Within the Divisions and Superorders . . . . . . . . . . . . . . 350
Division I. ....... ....... 350
Elopomorpha. .............. 350
Clupeomorpha ............... 350
Division II ........ ....... 350
Osteoglossomorpha . . . . . . . . . . . . . . . . . . . . . . . . . . 351
Division II I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
Protacanthopterygii ............. 352
Ostariophysi ............ ... ... .. . 352
Paracanthopterygii ....................... 352
Atherinomorpha. ..................... 353
Acanthopterygii ....................... 353
Relationships and Composition of Certain Superorders, Orders, and Suborders . . . 354
Division I. ......... .. ............ 354
Superorder Elopomorpha . . . . . . . . . . . . . . . . . . . . . . . . 354
Superorder Clupeomorpha ......... .............. 358
Division II ........................ 361
Superorder Osteoglossomorpha . . . . . . . . . . . . . . . . . . . . . 361
.

Division III ........................ 366

Superorder Protacanthopterygii ........ ........... . . . 366
Problematical Suborders of the Order Salmoniformes . . . . . . . . . . . 372
Suborder Stomiatoidei. . . . . . . . . . . . . . . . . . . . . . . 372
Suborder Alepocephaloidei. . . . . . . . . . . . . . . . . . . . 373
. .

Significance of Order Gonorynchiformes in History of Ostariophysan Fishes . . 374

Superorder Ostariophysi . . . . . . . . . . . . . . . . . . . . . . . . . 380
Order Cypriniformes . . . . . . . . . . . . . . . . . . . . . . . . . 382
Suborder Characoidei . . . . . . . . . . . . . . . . . . . . . . . . 383
Suborder Gymnotoidei . . . . . . . . . . . . . . . . . . . . . . . 383
Suborder Cyprinoidei . . . . . . . . . . . . . . . . . . . . . . . 384
Order Siluriformes . . . . . . . . . . . . . . . . . . . . . . . . . . 386
Superorder Paracanthopterygii .................... . . 387
Superorder Atherinomorpha . . . . . . . . . . . . . . . . . . . . . . . 390
Superorder Acanthopterygii . . . . . . . . . . . . . . . . . . . . . . . 390
PROVISIONAL OUTLINE CLASSIFICATION OF THE TELEOSTEAN FISHES. 393
BIBLIOGRAPHY..
......................... 436
INDEX TO NAMES IN PROVISIONAL OUTLINE CLASSIFICATION OF THE TELEOSTEAN FISHES 445

343
 INTRODUCTION
HISTORY
THE LATEST widely accepted general classi- lineages from diverse holostean ancestors in
fication of teleostean fishes is that of Berg the Mesozoic. Indeed, during the past 35 or
(1940), and in neither its second edition, 40 years, it has become generally recognized
edited by Svetovidov (Berg, 1955), nor its by paleo-ichthyologists that the holosteans
German translation (Berg, 1958), is the ar- themselves represent merely a stage or level
rangement of the teleosts materially altered. of organization into or through which nu-
Berg's teleostean groupings, like those of Jor- merous actinopterygian lines passed during
dan (1923), closely reflect the conclusions their evolution from separate stocks of Late
reached by Regan in a long series of brilliant Palaeozoic or Early Mesozoic palaeoniscoid
papers culminating in his brief general expo- derivatives. Woodward's paper was not
sition (Regan, 1929). Regan's teleostean pa- widely noted by students of living teleosts,
pers, in turn, were built upon the much ear- but Bertelsen and Marshall (1956), in dis-
lier foundations laid by Boulenger (1904) and cussing the mirapinnids, explicitly supported
Woodward (1901), and, in the nineteenth the view that different teleostean lineages
century, by Gunther, Cope (1871), and Gill have attained certain comparable grades of
(1872, 1893). In fact, except for relatively mi- organization. The idea of teleostean poly-
nor revisions, shifts, and splitting, most of phyletism was expressed by Bertin and
the major groups of living teleosts recognized Arambourg (1958, p. 2208) in their extensive
by Berg do not depart in any revolutionary account of the group. However, their treat-
way from those recognized 70 to 90 years ago ment of the living forms appeals to us as
by Gill. chiefly another reshuffling of long-recognized
In reviewing the literature on fish classifi- entities, improved here and there by Aram-
cation, we were greatly impressed by an bourg's extensive knowledge of fossil teleosts,
almost forgotten scheme proposed by Gar- but marred by an unfortunate lack of fa-
stang (1931). Garstang's classification was miliarity with many Recent groups as well as
surprisingly modern in concept and was in by such egregious errors as acceptance of Y.
effect a rebellion against the rigidity of a Le Danois' (1961) imprecise and unaccept-
taxonomy in which the compartments were able work on the tetraodontiforms.
artificially arranged and rounded off. In Gar- Thus we are left at the present day with no
stang's own words, a classification consisting general classification of teleostean fishes that
of " . . . an array of detached and isolated or- has utilized those modern concepts of phy-
ders, which conveys no explicit outline of the letic classification that have become com-
evolutional succession . . . " is wrong in that mon in the study of mammalian evolution
"Its very flatness and lack of relief is indeed (e.g., Simpson, 1945). Yet the feeling has
a misrepresentation of nature." been growing among our group and others
More than 20 years ago, Woodward (1942) that many of the most generally recognized
published a prophetic little paper on the be- teleostean orders are no more than catch-alls
ginnings of teleostean fishes, in which he ad- for separate lineages which have attained a
vanced the view that the Teleostei, long rec- comparable stage of specialization or com-
ognized as a natural, monophyletic group, in plexity (see particularly Gosline, 1960).
reality had evolved as a number of distinct
THEORY
The problems faced in attempting a "na- relationships must be inferred largely on the
tural" taxonomic classification of a large and basis of the morphology of living stocks. The
varied group of organisms have been widely question of "horizontal" versus "vertical"
discussed (see Rensch, 1960, and especially classifications in such groups becomes essen-
Simpson, 1961). The problems are greatest tially that of typological versus phyletic
when the meaningful part of the fossil record taxonomy, which Simpson (1960, pp. 46-66)
is relatively scanty, as it is in the teleosts, and has adequately discussed.
345
346 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
Teleostean classification, up to and includ- tensive study of living assemblages, are rarely
ing not only Berg's work, but also a very very full or precise and are seldom used ex-
large part of that of Bertin and Arambourg cept by those who wish to emend them.
(1958), has been arrived at primarily by However, unless one wishes to abandon the
methods that are essentially typological in principle that taxonomic classification should
nature-an attempt first to define orders and reflect what can be determined of phylogeny
other higher taxa and then to speculate upon (as some people do), taxa that are obviously
their origin, albeit in the light of the known polyphyletic must be broken up and a new
fossils. In the mammals, the preponderant classification must be adopted.
weight of the fossil over other evidence long The classification that we now propose is
ago forced mammalogists to the phyletic type based on an analysis of what we consider to
of classification. No such revolution in tele- be the predominant evolutionary trends in
ostean classification has occurred up to the the teleosts. By basing the definitions of
present day. groups solely on these trends, we have tried
Of the present authors, Myers has thought to free teleostean classification as much as
for a number of years that the varied "order" possible from the confining influences of ty-
of "isospondylous" or "clupeiform" fishes is pology.
a polyphyletic assemblage; Greenwood This classification is not intended to be de-
(1963) has already begun the demonstration finitive at any level. As originally conceived,
that such is true; while Rosen (1964 and other it was to be nothing more than a series of dis-
work) has begun the dismemberment of Re- cussions outlining some new evidence bearing
gan's "percomorphs," and the preliminary on fish relationships, some new thoughts
demonstration of the affinities of certain stemming from the reconsideration of old evi-
perchlike groups with the relatives of the dence, and a synopsis of certain outstanding
salmonoids. Moreoever, all of us (see espe- problems. As the work proceeded we realized
cially Weitzman, 1964, p. 154, and our dis- the need to fill in those areas with which we
cussion below on the gonorynchoid fishes) were not primarily concerned. Indirectly this
have more recently begun to think that the led to the gradual compilation of a list of fam-
ostariophysan fishes may be far older than ily-group and higher taxonomic names. This
was previously believed and contain separate list, with the subsequent addition of synony-
lineages running back to a generalized Meso- mic familial and ordinal names, is included
zoic teleostean. Our prime purpose, then, has and is intended merely as a nucleus for fur-
been to separate and point out what we be- ther search. A series of outline drawings illus-
lieve to have been the main and subsidiary trating each of the major families mentioned
phyletic lineages of teleosts and the often par- in the classification is appended.
allel or converging trends that characterize The families recognized, and their place-
the evolution of these lines. We have not been ment, follow Berg's (1940) arrangement but
especially interested in the definition of living with emendations based on works published
(or fossil) "groups" as such. Definitions of subsequently and on unpublished informa-
higher taxa, even those based on deep and ex- tion supplied by our colleagues.

TELEOSTEAN DIVERSITY AND AGE

Although the teleosteans are far from be- which the actinopterygians have flowered
ing the well-circumscribed group that Jo- into the largest (hence, by some definitions,
hannes Muller (1845) and his successors be- the most successful) of all major vertebrate
lieved them to be, they are at the present groups.
time well separated from the living holos- Their diversity is astounding. Estimates of
teans and chondrosteans. The term "tele- the number of living species vary from some-
ostean" (or "teleost") has meaning, even if it what under 20,000 up to 40,000. The facts
represents merely the final grand stage which that discovery of new species and genera is
so many diverse lines of actinopterygian (or still commonplace, and that new forms of
teleostome) fishes have attained, and within considerable evolutionary importance (e.g.,
 s f - P - z

BULLETIN AMER. Mus. NAT. HIST.

-s A

VOL. 131, PLATE 21

'm'4
c~~~~~~~~~
~ 1 XXE

C E

L. -
0 J a-)~~~~~~~/
(~~~~~~~~~~~~~
 °
BULLETIN AMER. Mus. NAT. HIST. VOL. 131, PLATE 22

C-t
Q

Q
CZ,
CT,

ci,

C*,)

9X4
Ut
£4

-4-
0

CL
Oa
W
BUlLETIN AMER. Mus. NAT. HIST. VOL. 131, PLATE 23

VCO

5.? >-
0

Cd

0-

05
*co)
Cd-
50

c 0
cd
44.0

443)

Co

Cd
Co

Sc
0
0

E
CO)

Co

Qct
0
to>

L.. OwC
0 Vl
4-

V)04-
C1)
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 347
Denticeps and many recently described deep- seas laid down during the Late Cretaceous in-
sea forms) are still being discovered at a sur- dicate that several lineages had by then at-
prising rate, demonstrate that we are further tained an organization similar to that of liv-
from a reasonably complete knowledge of liv- ing forms. This statement is especially true of
ing teleosts than we are of any other large, the elopoid and berycoid lines. However, the
non-piscine, vertebrate group. Bailey (1960) absence in known Cretaceous deposits of sev-
estimated the present total to be somewhat eral important lines of teleostean develop-
fewer than 17,000 species, and Myers (1958) ment (notably the salmonoids and ostario-
estimated that the eventual total number of physans, which give considerable evidence of
living species will approach or surpass 30,000. an age comparable to that of the elopoids)
The most numerous additions to the total again leads to the suspicion that much tele-
may be expected in the deep seas and in the ostean evolution was going on in Mesozoic
excessively rich fresh-water fauna of tropical fresh waters-evolution of which we as yet
America. have no trace.
Unlike mammals, of which most of the liv- By the Eocene, or possibly even the Pale-
ing orders had their origin in the Cenozoic ocene, teleostean marine shore-faunas bore a
and of which extremely few relicts of Meso- striking resemblance to modern assemblages,
zoic or early Cenozoic type persist, a consid- a fact that again is wholly unlike the situa-
erable number of living teleosts belong to tion in mammals. Since that time, a number
genera close to or perhaps even identical with of teleostean families appear to have under-
Eocene, Paleocene, or Cretaceous forms. Also gone comparatively little change.
unlike mammals, the number of known living Unfortunately, we know comparatively
genera and species far surpasses the number little about Paleogene fresh-water or deep-
of known fossils. For these reasons, a classifi- sea fishes in any part of the world. We may
cation based principally on the Recent tele- presume that ostariophysan types not greatly
osts, as this one is, has far more validity than different from living forms were in residence
would a classification based on Recent mam- on nearly all the continents, a presumption
mals, or the Recent forms of other large ver- based on some concrete fossil evidence. The
tebrate groups. appearance of undoubted bathypelagic stomi-
Fishes of teleostean type (Leptolepididae) atoid and myctophoid fishes in Early Neo-
first appear in the known fossil record in the gene records, in which these fishes are of
Middle Triassic. Some of these are so ad- wholly modern type, leads also to a presump-
vanced in the details of their structure that tion that they, too, arose much earlier.
we can speculate that the shift from the holos- We have not excluded fossils from consid-
tean to the teleostean level began much ear- eration, although we do not place them in our
lier in some forms. However, the dearth of formal classification. A number of important
Early Mesozoic fossils of teleostean type, ex- fossils are discussed in the expository com-
cept in marine Triassic and Jurassic beds in ments preceding our classification. Paleo-ich-
the area of the Tethys Sea, may be related to thyologists who deal extensively with teleos-
a fresh-water origin of many teleostean lines tean fossils are quite aware that the classifica-
in regions where fresh-water, fish-bearing de- tion of living teleosts must be understood be-
posits are rare or undiscovered. The many fore the fossil record can be properly inter-
teleosts known from the deposits of epeiric preted.
 NATURE OF MAJOR GROUPINGS ADOPTED
ENOUGH IS SAID above about the evolution- inantly fresh-water, radiations, the Osteo-
ary trends exhibited by different lineages to glossiformes and Mormyriformes. Neither of
indicate our belief that these lineages often these orders could possibly have been in-
pass through or arrive at similar levels or volved in the ancestry of other teleosteans.
stages of organization. We propose, for ex- By contrast with the other divisions, Divi-
ample, that separate evolutionary routes sion III contains the bulk of the living tele-
toward the acanthopterygian grade had been ostean fishes. There have evolved within this
traversed by relatively unrelated lines, and division several radiations leading to more
we also propose that the malacopterygian than one organizational level and to the dom-
level probably was attained polyphyletically inant groups of extant fresh-water and ma-
from holosteans of pholidophoroid type. rine species, namely, the Ostariophysi and
The following series of synopses and dis- the Acanthopterygii.
cussions outline the reasons for the new align- In the discussions below of the major struc-
ments given in the formal classification. The tural and developmental divisional trends,
principal innovations in this classification are where we outline the reasons for erecting the
the separation of the teleostean fishes into various new orders and superorders, a sep-
three divisions' and the realignment of taxa arate analysis is given of the suborders Sto-
among eight superorders (fig. 1). Various miatoidei and Alepocephaloidei of Division
smaller groups (suborders and families) are III because of their previous placement near
redistributed among orders, both new and groups here included in Division I. The ordi-
old. nal and subordinal composition of the super-
In our conception, each of the three divi- order Ostariophysi is also given in detail. The
sions represents a distinct phyletic lineage de- one superorder not discussed at length is the
rived from the holostean level of organiza- Atherinomorpha. It does not fall readily
tion. It is presumed that in Division I a prim- within the concepts of either of two adjoin-
itive elopiform ancestor has produced prin- ing groups, the Paracanthopterygii and Acan-
cipally the eels and eel-like fishes and perhaps thopterygii, although it includesforms at more
also the herring-like fishes. However, we or less the same organizational level as the
know of no evidence to rule out the possibil- fishes of those superorders. The main struc-
ity that the herrings and their allies had an tural and developmental characteristics of
independent origin from among the pholido- the Atheriniformes, the only contained order
phoroids. For the present, and because elopo- of the Atherinomorpha, were described in
morphs and clupeomorphs appear more some detail by Rosen (1964).
closely related to one another than to other Among the various living primitive groups
groups, we have followed a conventional of teleosteans, only elopids (Division I) and
alignment of these fishes. salmonids (Division III) are sufficiently gen-
In Division II there have evolved only two eralized to be suitable morphologically as ba-
series of unusually specialized, and predom- sal types for the major teleostean radiations.
Other primitive groups, for example, the clu-
1 Parenthetically, we should mention that the possi- peiforms and osteoglossiforms, are too spe-
bility of utilizing the methods of numerical taxonomy cialized for this role. We realize that elopids,
to help solve many of the major questions of relation- of all teleosts, possess the greatest assemblage
ships has occurred to us. Our reluctant conclusion has
been that the work of encoding (for computer tech- of holostean characters. At the same time, we
niques) the at times esoteric (and so heavily weighted) recognize that their larval and other special-
evidence that we have dealt with would not only have izations, and the absence of certain snout and
violated the weightless spirit of numerical taxonomy, jaw structures (see discussions below), put at
but perhaps have postponed presentation of our results
until our respective retirement ages. However, forthe least the living elopids off the main course of
benefit of our "numerical" colleagues, we have numbered teleostean evolution. Salmonids, on the other
our major divisions (I, II, III). hand, have none of these limitations and thus
348
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 349

FIG. 1. Diagram showing our conception of the evolutionary relationships of the princpal
groups of teleostean fishes. Uncertain relationships are shown by a broken line and question
mark.
350 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

seem better to fulfill the requirements of a sider them to be holosteans), that a common
morphotype that may have given rise to the ancestor of these two modern groups, if it ex-
major radiations of Division III. A question isted, was itself likely to have been an ad-
that naturally arises from this conclusion con- vanced type of holostean fish. If this view is
cerns the possibilities that elopids and sal- supported by paleontological evidence, as we
monids might have arisen polyphyletically think it ultimately will be, the salmonids and
from the holostean level or monophyletically elopids would represent separate attainments
from a single holostean or early teleostean en- of the teleostean grade and thus would be, ex
tity. Elopids are still so close to the holostean hypothesis, examples of polyphyletism, at
grade, however (indeed, some workers con- least at the teleostean level.

MAJOR TRENDS WITHIN THE DIVISIONS AND SUPERORDERS

DIVISION I line canals, hence the formation of a recessus
lateralis.
Fishes of ancestry at or near the holostean 11. The development of a leptocephalous larva.
level of organization in which each contained
group except for the eel-like fishes has one or ELOPOMORPHA
more very primitive members. 1. Principally marine fishes of diverse form,
Characteristic and often primitive trends most of the modern species eel-like.
include: 2. Gular plate in non-eel-like representatives.
1. The development, particularly in the com- 3. Branchiostegals usually very numerous.
pressed, silvery, marine fishes, of a short, broad, 4. Mesocoracoid arch present only in the non-
and arcuate maxilla equipped with large movable eel-like forms.
supramaxillae in association with a high coronoid 5. Hypurals, when present, on three or more
process onthe dentary and articular. centra.
2. The development of maxillary teeth that are 6. Ethmoidal commissure present or in modi-
seldom excluded from the gape, even partially. fied state in many groups.
3. The development of parasphenoid and 7. Opercular series often reduced or even
pterygoid teeth. absent.
4. The development in the basicranium of nu- 8. Larva, when known, a leptocephalus.
merous separate intraosseous passages for the CLUPEOMORPHA
parts of the fifth and seventh cranial nerves and
certain major blood vessels. 1. Silvery compressed fishes, usually marine,
5. The development of a full complement of with caducous scales.
intermuscular bones. 2. Branchiostegals numbering as high as 15,
6. Caudal fin, when present, with hypural sup- but usually fewer.
ports on one to four centra. 3. Intracranial diverticula of swim bladder
7. The development of a functional ductus forming bullae within the ear capsule.
pneumaticus. 4. Mesocoracoid arch invariably present.
8. The development of an otophysic connection' 5. Hypurals on one to three centra.
not involving the intercalation of bony elements. 6. Cephalic lateral-line canals extending over
9. The development of an ethmoidal commis- operculum; usually no lateral-line pores on trunk.
sure of the cephalic lateral-line system. 7. Recessus lateralis present.
10. The development of a confluence between
the preopercular and infraorbital cephalic lateral- DIVISION II
Fishes of ancestry at or near the holostean
1We define "otophysic connection" as the involve-
ment of the swim bladder with the otic region, usually
level of organization in which all contained
effected through bilateral prolongations of the swim members have retained numerous primitive
bladder which either come into superficial contact with characteristics of the jaw suspension and
the cranium in the region of the inner ear or penetrate shoulder girdle, and have developed com-
the cranial wall and form an intimate association with plexly ornamented scales.
the inner ear intracranially. Exceptionally the swim Characteristically primitive trends, and
bladder-ear connection may be effected through a chain
of ossicles linking the perilymphatic cavity with the some of the divisional specializations, in-
swim-bladder wall. clude:
1966 GREENWOOD, ROSEN, WEI TZMAN, MYERS: FISHES 351
1. The fusion of the premaxillae into a single the premaxillae firmly bound to the ethmo-
bone. vomerine region in all genera.
2. The development of a simple, well-toothed 3. Primary bite of mouth between parasphenoid
maxilla (edentulous in the Mormyriformes), gen- and glossohyal and basihyals.
erally contributing to the gape but partially ex- 4. Head of palatine without maxillary process.
cluded in a few genera. 5. Branchiostegals three to five, in two cases 11
3. The development of parasphenoid, glosso- and 13.
hyal, and pterygoid teeth. 6. Subtemporal fossae present in only a few
4. The fusion of various elements in the palato- genera.
pterygoid arch, the palatine and vomer fused in 7. Expansive suprascapulars in all except the
Mormyriformes. Osteoglossoidei.
5. The development of paired tendon bones 8. A lateral cranial foramen in most species
(uncalcified in Hiodon) on the second hypo- (Osteoglossoidei excepted).
branchial or second hypobranchial and basi- 9. Hypurals, in fishes with distinct caudal fins,
branchial, in all genera. reduced in number by fusion of the upper elements
6. The loss, in many species, of multiple intra- and, apparently, supported on at least two and a
osseous passages in the prootic bone for the fifth half centra in all.
and seventh cranial nerves and certain major
blood vessels. DIVISION III
7. A reduction, in two phyletic lines, of the
caudal fin and its confluence with the dorsal and Fishes mostly of distinctively teleostean
anal fins. level ancestry, only a single basal group hav-
8. The development, in those genera with a dis- ing obviously holostean affinities.
tinct caudal fin, of a supporting skeleton of which Characteristic trends include:
the elements are not readily homologized with 1. The lowering of the center of gravity and the
similar elements in fishes of other divisions. There
is a reduction in the number of hypural elements approximation of the center of buoyancy with the
which are apparently supported by two and a half center of mass.
centra in all except one genus (Hiodon), in which
2. The development of a large, frequently
three centra are involved and there is a full com- mobile premaxilla that completely or partially ex-
plement of hypurals. cludes the maxilla from the gape.
9. A reduction in size, or loss, of the subopercu- 3. The loss of maxillary teeth and functional
lum. supramaxillae.
10. The development of upper intermuscular 4. The loss of parasphenoid and pterygoid
bones only. teeth.
11. The development of a functional ductus 5. The development of an os pharyngeus in-
pneumaticus. ferior and an os pharyngeus superior and the de-
12. The development of an otophysic connec- velopment of mm. retractores arcua branchialia
tion not involving the intercalation of bony ele- attached to the third to sixth vertebrae.
ments in the young or adults of all groups except 6. The development in the basicranium of a
the Osteoglossoidei in which no connection exists. common passage (trigeminofacialis chamber) for
13. The development of distinctly separated the fifth and seventh nerves and orbital artery and
preopercular and infraorbital canals and the oc- head vein.
currence in a single genus (Pantodon) of a supra-
7. Loss of the supraorbital bone.
preopercular bone. 8. The reduction in size of the infraorbital
14. The development of somatic electric organs bones.
in one order. 9. The reduction in the number of scale bones
in the dorsicranium.
OSTEOGLOSSOMORPHA 10. The loss or reduction of certain temporal
fossae but the enlargement of the posttemporal
1. Fresh-water and predominantly tropical fossa and the loss of its roof.
fishes of extraordinarily diverse body form and 11. The covering of the posterior part of the
size, and including one form in which the pectorals dorsicranium by epaxial body muscles.
are greatly expanded so as to give the impression 12. The elevation of the pectoral fin base on the
of a flying fish. Most species are insectivorous or side.
piscivorous. 13. The forward migration of the pelvic girdle
2. Premaxillae ankylosed to form a single and its linkage with the pectoral girdle.
median bone in one order (Mormyriformes) and 14. The reduction in the number of pectoral
in one genus of the Osteoglossoidei (Pantodon); radials.
 352 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
15. The reduction in the number of vertebrae 19. Suprapreopercular (canal-bearing ossicle
and of pelvic and caudal fin rays. above uppermost part of preopercular canal)
16. Reduction of intermuscular bones. present in generalized representatives.
17. Reduction of the hypural bones to a single
unit on a terminal half-centrum. OSTARIOPHYSI
18. The varied specialization of caudal fin 1. Predominantly fresh-water fishes of world-
shape. wide distribution on the continents and adjoining
19. The development of an adipose fin in several archipelagoes, of extraordinarily diverse form and
primitive lines. habits, encompassing numerous well-toothed pred-
20. The development of fin spines and ctenoid atory and vegetarian types and toothless detritus
scales. and microphagous types, many of both categories
21. The development of an otophysic connec- with well-developed circumoral barbels.
tion involving the intercalation of bony elements. 2. Upper jaw protrusile in numerous species.
22. The disappearance of the ductus pneu- 3. Major trends toward reduction in number
maticus. (or absence) of jaw teeth.
23. The development of distinctly separated 4. Lower pharyngeal bones usually well de-
preopercular and infraorbital canals, hence the veloped.
frequent occurrence of a suprapreopercular bone. 5. Branchiostegals generally few in number but
24. The development along the trunk of a as many as 15 in some species.
ramus lateralis accessorius of the seventh nerve. 6. Pelvic fins abdominal.
7. Hypurals on one centrum.
PROTACANTHOPTERYGII 8. Fin spines present in numerous instances.
1. Predominantly slender, predatory fishes; 9. Scales present or absent, when present
many generalized and some specialized forms in cycloid in most instances, ctenoid in a few, and in
fresh water. certain forms replaced by dense, bony plates.
2. Photophores in oceanic representatives. 10. Adipose fin in many groups.
3. Widespread trend toward exclusion by pre- 11. Otophysic connections involving the inter-
maxillae of the maxillae from the gape. calation of bony elements in all.
4. Widespread trend toward the development 12. Swim bladder primitively subdivided, re-
of premaxillary processes. duced in many species.
5. Palatopremaxillary and ethmomaxillary liga- 13. Suprapreopercular (ossicle above upper-
ments present in numerous representatives. most part of preopercular canal) in numerous spe-
6. Upper jaw slightly protrusile in a few cases. cies.
7. Glossohyal teeth usually prominent.
8. Branchiostegals very numerous in many in- PARACANTHOPTERYGII
stances, reduced to two or three in some cases. 1. Mostly marine, stout, soft-bodied fishes in-
9. Hyoid and branchiostegal skeleton approach- habiting deep waters or when in shallow water
ing paracanthopterygian and acanthopterygian being nocturnal or occurring in cryptic habitats.
form. 2. Virtual loss of photophores.
10. Paired proethmoids present in many cases, 3. Feeding mechanism adapted for carnivorous
often simulating ascending premaxillary processes. diet in all species.
11. Few species with opercular spines or ser- 4. Ascending process of premaxilla often joined
rations. to premaxilla by flexible cartilage, or absent;
12. Mesocoracoid present in generalized lines premaxilla with an articular process in all cases,
only. and with a lateral (maxillary) process in most
13. Baudelot's ligament to first vertebra. cases.
14. Occasional trends for the pelvic fins to ad- 5. Ethmomaxillary and palatopremaxillary lig-
vance; pelvics commonly of more than six rays. aments well developed.
15. Occasional trends toward elevation of the 6. Upper jaw not protractile.
pectoral fin base on flank. 7. Mm. levator maxillae superioris well de-
16. Vertebrae commonly more than 24, pre- veloped, or modified and consolidated with part
caudal elements commonly 15 or more. of m. adductor mandibulae.
17. Hypurals on one to three centra, but a 8. Superficial division of m. adductor mandi-
basic acanthopterygian caudal skeleton developed bulae reduced or absent.
in some representatives and a paracanthoptery- 9. M. adductor arcus palatini covering floor of
gian type in others; caudal fin commonly with orbit.
more than 15 branched rays. 10. Ceratohyal and epihyal ankylosed.
18. Adipose fin present in most species. 11. Branchiostegals not exceeding six in num-
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 353
ber, the bladelike elements with an anteroproximal 4. Superficial division of m. adductor mandi-
prong in most species; the four bladelike elements bulae present and well developed, with a tendon
on the outer face of elevated part of hyoid bar, to the lower maxillary shaft.
the anterior hairlike elements on the inner side of 5. Upper and lower pharyngeal bones well de-
the depressed anterior section of ceratohyal. veloped, dentigerous, the upper bones consisting
12. Upper and lower pharyngeal bones well of a large plate made up of pharyngobranchials 3
developed and toothed. and 4 and smaller modified pharyngobranchial 2;
13. No subocular shelf on infraorbital bones. pharyngobranchial 1 present but obsolescent in
14. Extrascapular bones present, often forming only a few instances.
solid roof for posttemporal fossae. 6. Ceratohyal and epihyal joined together by
15. Parietals meeting in midline or closely ap- dorsal lamella.
proaching one another in most species and fre- 7. Branchiostegals four to 15 in number.
quently housing a posttemporal commissure of the 8. Mesethmoid usually bilaminar, invariably
cephalic lateral-line system. discoidal or scalelike.
16. Intercalar very extensive in numerous spe- 9. Infraorbital series reduced to two, rarely
cies. three, elements.
17. Mucous canals prominent on head of most 10. Opercular bones unarmed.
species. 11. Pectoral radials four in number, cuboidal,
18. Baudelot's ligament to first vertebra, or to recessed within excavation in scapulocoracoid
basicranium where first vertebra is fused to basioc- margin.
cipital. 12. Supracleithrum, when present, discoidal,
19. Modified epipleural ribs ("endocleithra") confined within dorsal tip of cleithrum.
from exoccipitals to cleithrum in several species. 13. Baudelot's ligament to basicranium.
20. Mesocoracoid absent; pectoral radials two 14. Pelvic girdle abdominal, subabdominal, or
to 13, often hourglass-shaped, very long, and ex- thoracic.
tending well beyond the scapulocoracoid margin. 15. Vertebral number high in most species, pre-
21. Pelvic fins thoracic, jugular, or mental in all caudal number modally 20.
but one species, with occasionally as many as 17 16. Caudal skeleton with two large hypural
rays. plates of opposite symmetry on terminal half-
22. Pleural ribs often reduced, frequently ab- centrum, with in no instance more than four
sent. hypurals, of which two are invariably broad and
23. Caudal skeleton, when present, with two fan-shaped.
large hypurals on separate vertebrae in most, or 17. Fin spines present or not.
the two fused together into a single unicentral 18. Ctenoid scales in relatively few species.
unit. 19. Numerous viviparous species, some with
24. Fin spines developed or not. unique encapsuled or unencapsuled spermato-
25. Ctenoid scales developed in some species. phores.
26. Swim bladder frequently subdivided and 20. In oviparous species, egg large, demersal,
connected by diverticulae to parapophyses of with adhesive filaments, and without oil globule.
precaudal vertebrae, in some instances an oto- 21. Embryo with heart displaced forward an-
physic connection involving the intercalation of terior to head.
bony elements.
27. Numerous species viviparous. ACANTHOPTERYGII
1. Fishes of extremely variable form and habits,
ATHERINOMORPHA principally in salt water and principally benthic
1. Generally small surface-feeding fishes, prin- and littoral.
cipally in fresh and brackish water, some marine, 2. Photophores very uncommon.
most fresh-water species with pronounced sec- 3. Feeding mechanisms extremely varied, per-
ondary sexual dimorphism in size, in color, and in mitting the utilization of numerous food sources.
fin shape and function; many species with bony -4. Upper jaw protractile in many species, with
external male genitalia developed from anal, pel- a premaxilla having ascending, articular, and
vic, or pectoral fin, or some combination of these. lateral (maxillary) processes.
2. Upper jaw protractile in many species, with- 5. Palatopremaxillary and ethmomaxillary lig-
out true ascending processes, and supported by a aments present, but in some cases modified.
foundation of loose connective tissue and a com- 6. Mm. levator maxillae superioris muscle ab-
plex maxillary process, without palatopremaxil- sent in all but one genus (Polymixia).
lary or ethmomaxillary ligaments. 7. Superficial division of the m. adductor man-
3. Mm. levator maxillae superioris absent. dibulae well developed.
 354 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 13 1
8. M. adductor arcus palatini usually confined exceeding four in number, often hourglass-shaped.
to posterior wall of orbit. 18. Pelvic fins, if present, thoracic or jugular in
9. Upper and lower pharyngeals well developed position, pectorals inserted high on the sides.
and toothed. 19. Pelvic fin typically consisting of a spine and
10. Hyoid bar with ankylosed ceratohyal and five articulated rays except in berycoids and a
epihyal; distal, depressed section of ceratohyal few other forms.
with large foramen in many cases; elevated proxi- 20. Pleural ribs usually well developed.
mal part of ceratohyal and epihyal with four 21. Vertebrae commonly numbering 24, with
bladelike branchiostegals, the hairlike anterior usually equal numbers of caudal and
branchiostegals, when present, on inner surface elements, except in some elongate andprecaudal in most
of depressed distal section of ceratohyal. fresh-water forms.
11. A subocular shelf present on the infraor- 22. Hypural bones virtually always emanating
bital series in numerous species. from a single centrum; when on two centra, the
12. Infraorbital bones frequently in contact hypurals no fewer than six in number, in no case
with preoperculum. formed as two hypural plates as in the Para-
13. Bones of head commonly with numerous canthopterygii .
pungent spines. 23. Caudal branched rays in most species
14. Opercular apparatus armed in many spe- 17 in more primitive members of the group. 15,
cies. 24. Fin spines present in most
15. Baudelot's ligament usually attached to 25. Ctenoid scales common. species.
basicranium, rarely (Polymixiidae and some Scor- 26. Presumably uniformly physoclistic.
paenidae) to first vertebra. 27. Otophysic connections rare, in no case in-
16. Supracleithrum strutlike, extending above volving the intercalation of bony elements.
cleithral tip in most members of the group. 28. Viviparity uncommon.
17. Mesocoracoid absent; pectoral radials not 29. Egg shape and buoyancy highly variable.

RELATIONSHIPS AND COMPOSITION OF CERTAIN SUPERORDERS,

ORDERS, AND SUBORDERS
DIVISION I character found in every group is the posses-
SUPERORDER ELOPOMORPHA sion of a leptocephalous larva. Larval stages
are, of course, not known for every genus
The Elopomorpha are an assemblage of ex- here included in the
tremely diverse types varying from a group tocephalous larvae areElopomorpha,
known for
but lep-
every order
of primitive fishes sometimes classified with and suborder.
the holostean fishes (the Elopiformes; see Traditionally, the Elopiformes are linked
Nybelin, 1957) to one group (the Sacco- with the
pharyngoidei) showing such considerable 1940), butclupeoid fishes (Regan, 1909; Berg,
in our opinion the relationship, if
modification of the syncranial architecture it
that it has been considered distantly related theexists,
is very distant and probably is at
to all other bony fishes (Tchernavin, 1947a, and pholidophoroid level. The elopomorphs
1947b). clupeomorphs differ in many syncranial
characters, of which particular reference may
It is difficult to discern over-all unifying be made to the absence of a
characters among the Elopomorpha, but cer- in the elopomorphs recessus lateralis
tain common intergroup characteristics can in that order of roofed (see p. 358), the presence
be recognized. For example, the ethmoidal of posttemporal fossae,
subtemporal fossae,
commissure in the mesethmoid of Elopi- tially of the tongue-parasphenoidal a primary bite essen-
formes and Notacanthiformes (in both in- of type, and
stances associated with canal-bearing rostral morphs parietals that meet medially. The elopo-
ossicles) and the several resemblances (but missure have which
a bone-enclosed ethmoidal com-
is associated laterally with
especially the swim-bladder bauplan) be- small,
canal-bearing,
tween the Anguilloidei and the Notacanthi- line, 1961). In Albula rostral ossicles (Gos-
formes (see Marshall, 1962). The Sacco- is an extension of the and Pterothrissus there
pharyngoidei and Anguilloidei are linked infraorbital lateral-line
through the nemichthyid-like eels. The sole canal onto the premaxilla. This character is
probably unique among the living teleoste-
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 355
ans. Certainly no bone-enclosed ethmoidal the "primitive" members of other orders or
canal is detectable in any clupeomorph fish, superorders, particularly those of Divisions
hence reinforcing the division between the II and III. In other words, we believe that,
elopiform and the clupeomorph fishes (see once the elopiform level of organization was
also Gosline, 1961, p. 35). Wohlfahrt (1937) reached, the group entered an evolutionary
described an ethmoidal commissure in Sar- cul de sac.
dina pilchardus, but the canal is entirely Two probable derivatives of the early elo-
superficial and not associated with an ossifi- pomorph stem are the eels (Anguilloidei) and
cation. the Notacanthiformes.
The caudal skeleton differs considerably in The notacanths share few characters with
the two groups (Hollister, 1936; Gosline, the other elopomorphs except in the posses-
1960; and this paper, pp. 358-359), except for sion of a well-developed and ossified ethmoi-
the primitive clupeomorph Denticeps which dal commissure which is basically similar to
approaches the condition of the elopiforms. that of the Elopiformes but differs from the
Denticeps and the elopiforms possess the most peculiar ethmoidal chamber of the Anguil-
primitive caudal fin skeletons known in liv- loidei (Allis, 1903; Gosline, 1961). The
ing teleosts. greatest resemblance in ethmoidal commis-
The characteristic otophysic connection of sure arrangement is between the Albuloidei
the clupeomorphs is described on page 358, and the Halosauridae; in both the canal is
where a comparison is also made with the hypertrophied and its bony walls are formed
only elopiform fish (Megalops) possessing an into an elaborate trellis. In body form and in
ear-swim-bladder linkage. In Megalops the most syncranial characters the Notacanthi-
cranial swim-bladder diverticula are lodged formes differ considerably from the Elopi-
within bullae formed principally from the in- formes. Basically, the notacanth jaws are of
tercalars of each side (Ridewood, 1904; de a primitive type, with distinct premaxillae
Beaufort, 1909). This contrasts with the and maxillae forming the border of the upper
typical clupeomorph condition in which jaw. No crossed ligaments are developed in
there are paired diverticula on each side, relation to the upper jaw. However, the mo-
which are lodged in bullae of the prootics and bile palatoquadrate arch and the relatively
pterotics. immobile hyoid arch of the notacanthiforms
Fairly marked differences in syncranial are distinct specializations. Again, the loss of
architecture occur between Elops and Meg- parasphenoidal teeth and of several elements
alops, on the one hand, and Albula and Pter- in the neurocranium (autopalatines, supra-
othrissus, on the other (see Ridewood, 1904a). orbitals, basisphenoid, orbitosphenoids, and
Albuli and Pterothrissus present a somewhat pleurosphenoids) are specializations. The
more specialized picture, with the maxilla pectoral arch, although attached to the skull
virtually excluded from the gape, the gular by a small or ligamentous posttemporal, has
plate (a substantial bone in Elops and Meg- lost the mesocoracoids and postcleithra which
alops) greatly reduced and poorly ossified, are present in elopiforms (but absent from
and the parasphenoidal dentition absent from the Anguilloidei). The body of notacanthi-
Pterothrissus. The elopiform lineage is an an- forms is decidedly anguilliform, and the cau-
cient one. Fossil elopiforms are found in Up- dal fin skeleton is greatly reduced or absent.
per Cretaceous deposits of Europe, Africa, As in the Elopiformes, the opercular appa-
and Asia; the Elops-Albula dichotomy was ratus of the Notacanthiformes is complete; it
established by at least Eocene times and is considerably modified in the halosaurs (see
probably earlier, since the genus Istieus (Up- Marshall, 1962, for a full discussion of this
per Cretaceous) shows strong affinity with important taxonomic point) in which the pre-
Pterothrissus. operculum is greatly expanded. In that pa-
The sum of morphological characters pre- per, Marshall also discussed in detail the na-
sented by the Elopiformes is, despite certain ture of the fins in the three families here
specializations of the Albuloidei, one of prim- united as the Notacanthiformes. From a gen-
itiveness. Yet it is difficult to derive from it eral consideration of their cranial anatomy
any of the structural grades characterizing (especially the opercular region), fin structure,
 356 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
and swim-bladder anatomy, Marshall (1962) fin) poses a particularly intriguing problem.
concluded that there are no grounds for main- Convergence seems less likely in this instance
taining two orders (Halosauriformes and because of the markedly different ways of
Notacanthiformes) as proposed by Berg life characterizing the two groups. Gosline's
(1940). views (1959) on factors molding the body
The Notacanthiformes are an ancient form and skull structure of the Anguilloidei,
group, at least as old as the Elopiformes. Fos- i.e., their burrowing and hole-haunting hab-
sil Halosauridae (Echidnocephalus, Enchelu- its, are particularly relevant in this connec-
rus) very similar to the extant Halosaurus oc- tion as is the fact that several eels are bathy-
cur in Upper Cretaceous deposits (Marck, pelagic but still retain the "typical eel" im-
1858, 1863; Woodward, 1901). Similarities print. (Resemblances in body form between
between these fossils and the extant Halo- the Anguilloidei and Saccopharyngoidei seem
saurus indicate that little morphological to fall into the same category.)
change has occurred since the Upper Creta- The Anguilloidei differ from the Elopi-
ceous. No fossil Notacanthidae are known. formes and Notacanthiformes in many ways.
The fossil Protonotacanthus (Upper Creta- The jaw structure is particularly divergent
ceous) is not a notacanthid and should per- (the premaxilla fused with the ethmoid, up-
haps be referred to the Protacanthopterygii per jaw bordered both by the maxilla and the
(Arambourg, 1954). Likewise, the Cretaceous ethmopremaxillary bloc), and in eels there is
Dercetidae, thought by Boulenger (1904) to a marked reduction in the size of the opercu-
be notancanthid fishes, were shown by Aram- lar elements. The reduction apparently is cor-
bourg (ibid.) to have myctophoid affinities. related with an increase in the importance of
Similarities between the Anguilloidei the branchiostegal rays as the supporting
(Anguilliformes) and the Notacanthiformes, skeleton for the opercular membrane. (In
apart from the leptocephalous larva,' center Cyema, there are no branchiostegal rays and,
principally on a fundamentally similar and apparently, no opercular elements.) As Gos-
otherwise unique bauplan for swim-bladder line (1959) has shown, this modification of
structure and organization (Marshall, 1962). the gill covers is correlated with a peculiar
Marshall is of the opinion that " .. . the re- pumping mechanism for the branchial cur-
semblances are detailed enough to suggest rent. To be effective, such a pumping device
that the Heteromi and Apodes evolved from demands that the mouth and gill chambers
a common ancestor, which must have been an be widely separated. The posterior displace-
isospondylous kind of teleost." He also con- ment of the gills in eels has also resulted in
sidered a number of other similarities, includ- the posterior displacement of the pectoral
ing the absence of oviducts, the long, many- girdle, which thereby loses its connection
rayed anal fin, the abdominally situated, with the skull.
eight- to 10-rayed pelvic fins (comparing As in the elopiforms and notacanthiforms,
these with the eight-rayed pelvics of the fos- the parietals of the anguilloids meet medially
sil eel Anguillavus), the medioparietal skull, (and may even fuse). A supraoccipital is usu-
and the development of "spectacles" (a clear ally developed, but it is wanting in the Seri-
patch of head skin) over the eyes. Many of vomeridae, Nemichthyidae, and Cyemidae.
these characters require further study before In general the neurocranium is well ossified
convergence can be overruled (for example, and forms a compact wedge (the families
the oviducts are absent from some members noted above being exceptions). The hyo-
of the Salmoniformes, and the eel-like perci- palatine arch shows reductional trends, in
forms of the family Mastacembelidae have no case consisting of more than three ele-
"spectacles" and a long anal fin). The anguil- ments-hyomandibular, quadrate, and pal-
loid-notacanth similarity in body form (espe- atopterygoid. The last is often reduced to
cially the extreme modification of the caudal a thin lamina and is not infrequently absent.
1 We are indebted to Mr. Harrisson for information, The hyomandibular and quadrate have an
in press, about the halosaur leptocephalus he has recently intimate, interlocking union; a symplectic
discovered. A second halosaurid leptocephalus has been does not develop.
described by Mead (1965). Allis (1903) has described a peculiar me-
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 3S7
dian sensory chamber in the ethmoid of cer- [see Trewavas, 1933] but one that requires
tain eels. Since the chamber is not linked further research on adults and larvae of both
with the supraorbital lateral-line canal, its groups); loss of branchiostegal rays (cf.
homology with a true ethmoidal comnmissure Cyema); the marked anterior extension of the
is uncertain. However, the resemblance be- pterotics; and the development of mm. ab-
tween the chamber and the ethmoidal canal ductores mandibulae as the principal muscles
of Albula (also without supraorbital con- for opening the mouth (again, cf. Cyemza
nections) may be noted. Perhaps the an- [Trewavas, 1933]). The Anguilloidei and
guilloid ethmoidal cha-mber should be con- Saccopharyngoidei are also similar in having
sidered in the light of the rather profound the branchial skeIeton displaced posteriorly
modifications that have taken place in the and the pectoral arch free from the skull
evolution of the anterior region of the skull (mesocoracoid absent in both). In structural
in these fishes. detail the branchial apparatus of the two
From the characters discussed above, it groups is very different (Tchernavin, 1947a,
will be obvious that there are some, and 1947b), and the posterior displacement of the
apparently fundamental, resemblances gills and girdle in the Saccopharyngoidei is
among the Elopiformes, Notacanthiformes, probably attributable to the extreme de-
and anguilloid Anguilliformes. On the other velopment of the pharynx and not, as in the
hand, these groups have diverged greatly in case of the eels, primarily to the displacement
other characters. Since undoubted, if prim- of the branchial apparatus.
itive, Anguilloidei also occur in the Cre- The saccopharyngoids have a peculiar
taceous (Urenchelys [Woodward, 1901] and arrangement of the cephalic lateral-line
A nguillavus [Hay, 1903]) this divergence is system (the mandibular branch is absent),
of long standing. and the lateral-line organs are carried on
Extreme specialization of the jaws, ante- papillae similar to those found in the Cyein-
rior vertebrae, and neurocranium make it idae among the true eels (a condition that
difficult to assess the relationships of the occurs also in the ceratioid anglerfishes; see
Saccopharyngoidei. In most classifications Trewavas, 1933).
(see Bertin and Arambourg, 1958) these Jaw structure in the eels and saccopharyn-
bathypelagic fishes are included with or goids is very different. Indeed, the homology
placed near the eels. Regan (1912b), how- of the upper jaw bones in the Saccopharyn-
ever, believed them to constitute an order goidei is still doubtful. In adults, as Tcher-
(Lyomeri) structurally quite unlike the eels, navin (1947a) has shown, the upper jaw has
an order that " . . .may well have been de- the morphological relationships of a ptery-
rived from Iniomi such as the Synodontidae." goid arch. But, in the larva (Orton, 1963),
This conclusion stands in sharp contrast there is an upper jaw which is morphologi-
to Regan's earlier views (1909) in which cally normal, that is, in the maxillary posi-
he allied the eels and saccopharyngoids. The tion.
most extreme view is that of Tchernavin The Eurypharyngidae are unique among
(1947b), who believed that these fishes were living teleosts in having five holobranchs and
distantly related to all other bony fishes. In six visceral clefts.
our opinion, the relationship of the sacco- Whereas differences between the Sacco-
pharyngoids is with the Anguilloidei. Besides pharyngoidei and the Anguilloidei are several
the presence of a leptocephalous larva sev- and marked, nothing in these differences
eral other characters suggest an anguilloid suggests a relationship with other groups.
ancestry. Among these may be mentioned the With the exception of certain vertebral
great reduction or loss of the palatopterygoid characters, the supposed resemblances be-
arch (but see below); union of the hyoman- tween the Saccopharyngoidei and the Syn-
dibular and quadrate (no symplectic); loss of odontidae (Myctophoidei), which Regan
opercular elements (at least in the adult; see (1912b) listed, could be attributed to con-
Orton, 1963); absence of ventral hyoid and vergence correlated with the backwardly
branchial elements, at least in adults (a con- directed, upper-jaw suspension that occurs
dition comparable with that of adult Cyema in both groups. The resemblances between
 358 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
the saccopharyngoids and anguilloids are Papyrocranus, however, some intracranial
less likely to be the result of convergence. invasion has taken place, but this invasion
Within the Anguilloidei, many saccopharyn- differs in detail from that found in the Clupe-
goid characteristics are present in the Cyem- omorpha. In the Mormyriformes the vesicles
idae and Nemichthyidae. are intracranial but are situated in yet an-
The Monognathidae are an as yet imper- other part of the skull, are not encapsuled,
fectly known group (Bertin and Arambourg, and have lost their connection with the swim
1958), but one that nonetheless exhibits sev- bladder in adults.
eral characteristics of the Saccopharyngoidei. Three remarkably constant features in the
The larva is, however, unknown. Bertin and skull of clupeomorphs are the temporal fo-
Arambourg (op. cit.) placed the family in a ramina (bordered by the frontals and pa-
distinct order related to the saccopharyn- rietals), the pre-epiotic fossae (bounded by
goids, but the many resemblances to the the parietal, pterotic, and epiotic of each
latter do not seem to warrant such a divi- side), and the auditory fenestrae (surrounded
sion. by the prootics, exoccipitals, and basioc-
There is no fossil record for the Sacco- cipitals). The relative sizes of these apertures
pharyngoidei. show considerable interspecific variation, and
SUPERORDER CLUPEOMORPHA from some genera one or, more rarely, two
may be absent; in most cases all three open-
Despite the varied habitus of fishes in- ings are present. A fourth and invariable
cluded in this superorder, the Clupeomorpha feature of the clupeomorph neurocranium is
are one of the most clearly defined aggre- the recessus lateralis, a chamber developed
gates of living teleosts. Three character com- on the medial aspect of the pterotic and
plexes serve sharply to distinguish these separated from the perilymph cavity of the
fishes, viz., the type of ear-swim-bladder inner ear by a thin membrane. The supraor-
connection, the architecture of the neurocran- bital, infraorbital, preopercular, and pterotic
ium (especially the relationship of certain lateral-line canals all open into this chamber.
lateral-line canals to one another and to the Perhaps concomitant with the development
ear), and the supporting skeleton of the of the recessus is the fact that the open-
caudal fin. ings of the infraorbital and preopercular
In all living clupeomorphs there is an inti- canals into it are very close together; in
mate intracranial connection between the some instances the openings are virtually
swim bladder and the inner ear. With very
few exceptions the swim-bladder diver- confluent, being at most separated by a nar-
ticulum of each side divides within the skull row bar of bone. Externally, the presence of
to form two large vesicles which are lodged this confluence is indicated in all clupeo-
within ossified bullae of the prootic and morphs by the way in which the upper parts
of the two canals approach each other. As
pterotic bones. The connection between these far as we can determine, the recessus later-
vesicles and the ear have been described in de- alis is a unique feature of the Clupeomorpha.
tail by Wohlfahrt (1936). In the exceptional With the exception of the Denticipitoidei,
genera (e.g., Sprattus), only the prootic ves- the caudal fin skeleton is remarkably
icle is developed. This unity of swim-bladder similar
bauplan in the Clupeidae, Chirocentridae, in all clupeomorphs and is of a type imme-
and Denticipitidae led Marshall (1962) to diately recognizable by several diagnostic
suggest that the families should be united in features (Gosline, 1960; Hollister, 1936). One
a distinct division of the suborder Clupeoidei. of these features is the way in which the sec-
The clupeomorph ear-swim-bladder con- ond hypural lacks a basal articulation with
the "urostyle," from which it is
nection is unlike that occurring in any other
group of fishes. In Megalops, for example, by a distinct gap. There is always separated
a certain
the single vesicle of each side is lodged with- degree of consolidation manifest in the caudal
skeleton; the urostyle is formed by the
in an intercalary vesicle, whereas in the Not-
opteroidei (see p. 362) the greater part of the of the terminal vertebra and the firstfusionuro-
vesicle is extracranial. In the notopterid neural. In some genera the caudal fin support
is unicentral, but in others a distinct centrum
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 359
is preserved posterior to that which carries The Denticipitoidei are known only from a
the first or lowermost hypural. single extant and monotypic genus in certain
The Denticipitoidei have a caudal fin skel- west African fresh waters, and a very closely
eton of an extremely primitive type. Three or related fossil from east Africa (Clausen, 1959;
possibly four centra are involved, the second Greenwood, 1960). Superficially these fishes
hypural (together with the third) is artic- have a very clupeoid appearance, but they
ulated with a centrum, the first uroneural ex- are immediately distinguished from the Clu-
tends to the centrum of the first hypural but peoidei by the presence of denticles on all the
is not fused with it (apparently only one roofing bones of the skull and even on some
uroneural is present), and small neural arches trunk scales. Anatomically the suborder is
are present on the two centra which carry the distinguished by the primitive caudal fin skel-
first three hypurals. In certain respects (es- eton (see above) and by several characters in
pecially the high number of discrete centra the syncranium (see Clausen, 1959; Green-
involved, and because the second hypural has wood, 1960, 1963, and 1965) of which may
not lost its basal articulation) the caudal skel- be noted the lateral-line system with its dis-
eton of the denticipitoids resembles that of an tinct frontal branch, the enlarged, gutter-
early larval clupeid (see Hollister, 1936). like, and medially contiguous nasals, and the
However, in its general organization, it shows general morphology of the skull, particularly
a degree of differentiation more comparable its smooth, almost larval, contours. Never-
with that of a larval clupeid in a later stage of theless, these peculiar little fishes have the
differentiation, that is, when the second hy- clupeomorph characters of a recessus later-
pural has lost its articulation with a centrum. alis, prootic and pterotic bullae for the swim-
Other characters of widespread occurrence bladder vesicles, a temporal foramen, and ex-
in the Clupeomorpha as a whole are: the vir- tensive lateral-line canals on the operculum.
tural absence of lateral-line pores on the Considering, especially, the nature of the cau-
trunk (except in the Denticipitoidei in which dal fin skeleton, the morphology of the neu-
a complete line is developed), the presence of rocranium, the structure of the jaws, and the
radiating cutaneous canals on the opercular presence of a complete trunk lateral line, we
bones (the canals open into the lateral-line believe that Denticeps is the most primitive
system but are themselves without neuro- living clupeomorph. The peculiar cephalic
masts), absence of parasphenoid teeth, no denticles and the absence of a supramaxilla
posttemporal or subtemporal fossae, and the are specializations, but these in no way ob-
ievelopment of keeled scutes along the ven- scure the generally primitive level of organi-
tral midline of the abdomen. zation demonstrated in so many other char-
The shape of the head and of the jaws acters.
among clupeomorph fishes is variable, al- Traditionally, the fishes here grouped
though usually the mouth is terminal and in the Clupeomorpha are allied with Elops,
moderately large and the dentary has a high Megalops, and Albula. Berg (1940), for in-
-~oronoid process. Most clupeomorphs are mi- stance, places all three genera in one subor-
-rophagous, and there is a marked tendency der. On the basis of the three major charac-
toward a reduction in oral and buccal denti- ters discussed above (otophysic connections,
tion, and a trend toward the development of neurocranial architecture, recessus lateralis),
suprabranchial organs associated with the we think that the Clupeomorpha are not
Fourth and fifth gill arches. An exception to closely allied to the elopoids and albuloids.
this generalization is, of course, the preda- Yet another difference between the two
-.eous Chirocentrus, which has large, fang- groups is the occurrence in the elopoid-albu-
Like teeth in the jaws. In no clupeomorph are loid aggregate of a leptocephalous larva.
the jaws protrusile. Ecologically, the super- In most respects the albuloids and elopoids
:)rder may be considered one of characteris- do not provide a suitable type of organiza-
tically marine, pelagic, and shoaling fishes. tion from which to derive the Clupeomorpha,
rhere are, however, numerous fresh-water especially the rather specialized clupeoids,
representatives, particularly in tropical Af- nor is it possible to derive the Denticipitoidei
rica and India. from the elopomorphs or vice versa. But the
 360 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VO L. 131
denticipitoid grade (apart from the loss of tion gathered from the literature, certainly
supramaxillae and the presence of cephalic suggests that this genus, and Pachythrissops
denticles) could well be basic to that of the (sensu Nybelin, 1964), differ from the clupeo-
clupeoids. The fundamental question is, Can morphs in many ways.
the proto-elopomorphs and proto-denticipi- The Clupavidae are an extinct family
toids be derived from a common stem? At (Lower to Middle Cretaceous) of particular in-
present the question cannot be answered. In terest. Currently, the clupavids are accepted
essence an answer first requires the elucida- as clupeoids having some affinity with the
tion of how the clupeoid type of otophysic round-herrings (Dussumierinae). Recently,
connection and the recessus lateralis were Whitehead (1963) has challenged this sup-
evolved. These two features, and perhaps also posed relationship and has suggested that the
the characteristically clupeomorph cranial clupavids may be " . . . between the elopoids
apertures, may be functionally interrelated. and Chirocentridae." Judging from Aram-
From evidence currently available from both bourg's reconstruction and
living and fossil fishes of the elopomorph and (Arambourg, 1954), we find figured material
difficulty in ac-
clupeomorph types, there is even the sugges- cepting either interpretation and suggest
tion that the two groups had a long separate third. The caudal fin skeleton of the Clu-a
history. Indeed, we have even given serious pavidae is, in many respects, clupeoid. That
consideration to the possibility that the Clu- is, a distinct "urostyle" is
peomorpha should be recognized as a distinct major hypurals (excludingdeveloped, and the
the first) articu-
division of teleosts. late with it. But the second
Although fishes known only from fossils ed and not free basally as hypural
is includ-
it is in all living
are not under revision in this paper, some clupeoids. The relationship of the infraorbital
comment is needed on those extinct forms and preopercular lateral-line canals in the
usually classified with the clupeoids. Clupavidae differs from that of all living clu-
Several extinct genera (e.g., Ichthyodectes, peomorphs: they do not approach each other
Xiphactinus, and Thrissops) are supposed to proximally and have widely separated open-
have some relationship with the Chirocen- ings into the supraorbital
tridae, a family showing several "primitive" arrangement suggests that canal. This canal
no recessus later-
characters. In our opinion, such belief is pre- alis was developed in these fishes.
mature, and the fossil material should be re- of the typical Also, none
viewed with regard to clupeomorph charac- has been clupeomorph cranial openings
ters described here. Thus we cannot agree in these characters described in the Clupavidae. Thus,
with Saint-Seine's (1949) or Bardack's (1964) considered as like thethe clupavids cannot be
placement of these and other genera in the their resemblance to this Clupeomorpha, despite
Chirocentridae. Indeed, our own observa- eral facies and, especially,group in their gen-
tions, coupled with the characters listed by phology. in their jaw mor-
Bardack (ibid.), suggest that members of his the Regrettably, no data are available on
Spathodacylus-Xip actinus line (including dae, endocranial morphology of the Clupavi-
so we cannot tell whether or not
the genera Ichthyodectes and Xiphactinus prootic
[= Portheus] which we have studied) show andInpterotic bullae are present.
certain affinities with the Osteoglossomorpha. the many respects, the Clupavidae resemble
Of particular significance in this connection Leptolepididae, a family of uncertain
position
are the basipterygoid process, the dentition, and known(sometimes considered holostean)
the presence of a subtemporal fossa like that to from deposits of Middle Triassic
in Osteoglossum, and the medially united There Upper Jurassic age (see also Rayner, 1936).
is nothing in
parietals. We suggest that these fishes may lepidids that could the structure of the lepto-
represent an early offshoot from the osteo- militate against their be-
glossid line; at present we are uncertain ing ancestral to the Clupavidae. We therefore
that a leptolepidid-clupavid phyletic
about the placement of those extinct genera suggestline was independently derived from
which Bardack (ibid.) put in his Thrissops- the
pholidophoroid stem and represents a con-
Chirocentrus line. Our detailed examination of vergent
a Thrissops species, together with informa- evolutionary trend toward that of
the clupeomorphs. If any common ancestry
1966 GREENWOOD, ROSEN, WEIrZMAN, MYERS: FISHES 361
for the two lines is sought, it should be looked parasphenoid and the entopterygoid (Osteo-
for among the pholidophoroid holosteans. glossoidei) or through fusion of the palatines
with the vomer (Mormyriformes). In most
DIVISION II Osteoglossomorpha, the maxillae are well
toothed, they form the greater part of the up-
SUPERORDER OSTEDGLOSSOMORPHA per jaw, and they are aligned in overlapping
'rhe extant Osteoglossoinorpha. are fresh- tandem with the premaxillae. The maxilla
water fishes confined, with one exception, to and premaxilla arefirmlybound together and
the tropical regions of South America, Africa, have restricted mobility; only in the Mor-
southeast Asia, New Guinea, and Australia. rnyriformes are the maxillae edentulous, rel-
The exceptional genus, ffiodon, occurs in atively motile, and partially excluded from
North Arnerica. Undisputed osteoglosso- the gape. In that order, the maxillae are ar-
rnorphs occur as fossils in fresh-water Eocene ticulated with the ethmoid, behind and above
deposits of Sumatra, Australia, and North the premaxillae. Supramaxillae are wanting
America; the supposed Osteoglossidae from in all osteoglossomorphs. The premnaxillae are
the Eocene of Europe and North Africa re- firmly bound to the skull and have appar-
quire, in our opinion, further study before ently very restricted motility. In Morrnyri-
their affinities can be deterrined. formes and in Pantodon, alone of the Osteo-
Despite the considerable range of body glossiformes, the prernaxillae are fused; they
form and jaw structure in members of this su- are closely apposed medially in Notopteroidei
perorder, the group is readily distinguished and in most Osteoglossoidei.
by two characters comrnon to all members, The second "group character" is the devel-
and by a number of other characteristics with oprnent of paired, usually ossified, rods at the
a rrosaic pattern of distribution anmong its base of the second gill arch in all Osteoglosso-
rmembers. rnorpha. These tendon bones develop in asso-
Nll Osteoglossomorpha have the primary ciation with the second hypobranchial (Os-
bite between the parasphenoid and the teoglossoidei and Mormyriformes) or the sec-
tongue (basihyals and glossohyal), the bones ond basibranchial (Notopteroidei; they are
involved being well toothed. In one genus uncalcified in Hiodont).
only, the osteoglossid Cupqisudis (=JIetero- The hypobranchial skeleton is variously
tis), is the prirnacy of the parasphenoid- developed in the superorder (see Ridewood,
tongue bite reduced by the absernce of teeth 1904b, 1905a). Unossified first basibranchials
on the parasphenoid and by a great reduction are a feature of Pantodon and Clupisudis (Os-
in the toothed area of the tongue. T'his loss teoglossoidei) and in the Notopteridae among
apparently is a secondary one because Clu- the Notopteroidei.
pisudis shows many of the other characters Characters of general occurrence, besides
associated with the typical condition. That the presence of paired subbranchial tendon
Clupisudis is the only microphagous osteo- bones, and the parasphenoid-tongue primary
glossomorph (the others are all predaceous) bite, are the fused parapophyses and the ab-
is undoubtedly significant, as is the develop- sence of lower intermuscular bones. The epi-
rnent in this genus alone amDng the osteo- neurals are confined to a few anterior verte-
glossomorphs of suprabranchial helices asso- brae in many cases.
ciated with the fourth and fifth gill arches The caudal fin skeleton is of particular in-
(d'Aubenton, 1955). Probably associated terest. In two families, the Notopteridae (Os-
with a primary parasphenoid-tongue bite is teoglossiformes) and the Gymnarchidae
(Mormyriformes), it is greatly reduced and
the fact that the maxillae show a greatly sin- there is no distinct caudal fin. All other mem-
plilied association with the palatines. There bers of the Osteoglossomorpha have a dis-
is no process for articulation between the tinct caudal fin, and in all except the Hiodon-
rnaxilla and palatine, the latter ending ante- tidae (Notopteroidei) there is a great similar-
riorly in a simple point. When bony articula- ity in its supporting skeleton. An outstanding
tions are developed between the skull and the characteristic is the considerable degree of
palatopterygoid they are either betveen the consolidation among the upper hypurals
362 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
which are also fused with the uroneurals and glossoidei and some Notopteroidei (Notop-
apparently with a vertebral centrum (Gos- terus and Xenomystus) but is an extensive,
line, 1960 and 1961). Attempts to determine shieldlike bone in other notopteroids (Papy-
the homologies of these elements raise several rocranus and, especially, Hiodon) and in all
difficulties, but, whatever solution is accepted, Mormyriformes. The parasphenoid is gener-
the caudal fin skeleton of the Osteoglosso- ally toothed but is edentulous in Clupisudis
morpha is quite distinctive and unlike that among the Osteoglossoidei and in some mor-
of any other teleost (see Gosline, 1960). The myriform species.
trend toward fusion shown by the Osteo- Apart from the Osteoglossoidei, all mem-
glossoidei and all Mormyriformes with dis- bers of the Osteoglossomorpha have, at some
tinct caudal fins is carried to extremes in the stage in ontogeny or throughout life, a con-
Notopteroidei and in the Gymnarchidae nection between the otic region and the swim
(Mormyriformes) which do not have an ex- bladder effected through a pair of forwardly
ternally distinct caudal fin. By contrast, the directed diverticula of the swim bladder. In
Hiodontidae (Notopteroidei) have an ex- the Notopteroidei, the connection persists
tremely primitive caudal skeleton, with a full throughout life, the swim-bladder diverticula
complement of hypurals and uroneurals being closely applied to the lateral aspects of
supported on three centra, the two anterior the otic region; an auditory fenestra is devel-
of which retain neural arches and spines. All oped (see Greenwood, 1963; Dehadrai, 1957).
osteoglossomorphs with a distinct caudal fin In one genus (Papyrocranus) extensive intra-
have 19 or fewer principal fin rays; one cranial swim-bladder diverticula are also
(Osteoglossum) has only 10. present (Greenwood, 1963). In Mormyri-
The pelvic fins are abdominal in all genera formes the otophysic connection is broken
except Pacntodon, in which they are thoracic; during ontogeny, but two vesicles remain
however, there is no connection between the lodged in each lateral cranial foramen or
pelvic and pectoral girdles. fossa (in Gymnarchidae). This fossa is an ap-
All members of the Osteoglossiformes, ex- erture bordered by the pterotic, epiotic, and
cept Pantodon, have characteristically orna- exoccipital, and is roofed by the extrascapu-
mented scales in which the radii form a coarse lar. The lateral cranial foramen or its homo-
and irregularly reticulated pattern, both api- logue (Greenwood, 1963) is present in the
cally and basally; no reticulum is developed Notopteroidei and in the Mormyriformes,
in the scales of Pantodon. A similar reticulum but no such aperture is present in the Osteo-
occurs in the scales of the Mormyriformes, glossoidei, thus suggesting an early dichot-
but it is confined to the apical area. Cockerell omy in the basic osteoglossomorph stem as-
(1910a, 1910b) noted that in Mormyriformes sociated with the evolution of an otophysic
the circulae are moniliform, but in the Osteo- connection. In no species is the otophysic con-
glossidae they are continuous. However, in nection like that found in the Clupeomorpha
the Eocene fossil osteoglossid Phareodus an (see p. 358), in which distinct intracranial os-
intermediate condition is found. The scales seous bullae are developed in association with
of the Notopteroidei do not resemble either the inner ear. Nor is the connection like that
the osteoglossid or the mormyriform types. of Megalops (Elopomorpha), in which the
Characters of the syncranium have a mo- swim-bladder diverticulum of each side is
saic distribution within the superorder. lodged in bullae formed from the intercalar
Most osteoglossomorphs have an orbito- laterally and the prootic, exoccipital, and
sphenoid (Osteoglossum and Scleropages of the basioccipital medially.
Osteoglossidae are exceptional); a basisphe- The inner wall of the swim bladder is
noid is present in some (Notopteroidei and cer- partly alveolate in Arapaima and Clupisudis
tain Mormyriformes); posttemporal fossae (Osteoglossidae), completely so in Pantodon-
occur only in the Osteoglossidae (except in tidae (Nysten, 1962), smooth in other mem-
Pantodon) as do subtemporal fossae (the bers of the Osteoglossidae and in all Mor-
fossa being more of a groove in Clupisudis myriformes except the Gymnarchidae in
and Pantodon). An extrascapular is present which it is alveolate. The abdominal portion
in all groups; it is a small bone in the Osteo- of the swim bladder in the Notopteridae is
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 363
completely subdivided (Dehadrai, 1957; ond and third suborbital bones. A subocular
Greenwood, 1963) but is simple and tubular, shelf is elsewhere found only in certain living
with smooth walls, in the Hiodontidae. In the and extinct Acanthopterygii.
Notopteridae, there is a considerable postab- A number of unique and interrelated char-
dominal extension of the swim bladder acters are found in the Mormyriformes, for
formed by paired diverticula lying on each example, the caudally situated electric organs
side of the vertebral column; ventrally each (modified body muscles), the very large cere-
diverticulum is produced into a number of bellum, and peculiarly modified neuromasts
viliform caeca. The notopterid condition is (so-called mormyromasts); the epidermis has
also seen in Osteoglossum, whereas in Clu- a peculiar histological structure (Franz,
pisudis there is a short postabdominal exten- 1920). Above and below each electric organ
sion contained within the haemal arches. In there is a thin, tendinous bone (bones of
all other Osteoglossomorpha there is no post- Gemminger) of unknown function.
abdominal extension. Ridewood (1905a) gave a summary of the
Throughout the Osteoglossomorpha there early taxonomic history of the Osteoglossi-
is a tendency for certain elements of the dae. From his account it is clear that several
palatopterygoid arch to fuse. In the Osteo- authors tended to associate this family with
glossoidei the ectopterygoid and palatine are the Hiodontidae, Notopteridae, and Mor-
fused, as they are in the Notopteroidei, al- myridae. Dissenting opinions were those of
though in Hiodon these elements are separate Woodward and Boulenger who related the
but firmly joined. In the Mormyriformes the osteoglossids to the Albulidae, a view en-
ectopterygoid, entopterygoid, and palatine dorsed by Ridewood himself (1905a), but one
are fused, and the latter is also fused with the that is difficult to substantiate when the char-
vomer. Ectopterygoidal, entopterygoidal, acters discussed above are taken into ac-
and palatine teeth are of common occurrence, count. Ridewood (1904b) also summarized the
although in some members of all groups cer- history of the Mormyridae, Notopteridae,
tain or all of these bones may be toothless. and Hiodontidae. On the basis of his cranio-
All, for example, are toothless in the Mor- logical studies Ridewood concluded that
myriformes. " . . . the families Mormyridae, Notopteri-
The Osteoglossoidei alone are characterized dae and Hiodontidae, though more closely re-
by an articulation between a ventrolaterally lated inter se than is either family with any
directed peg of the parasphenoid and the en- other family of Malacopterygian fishes are
topterygoid. No other group of extant tele- not more intimately related with one another
osts has such a joint. than was previously assumed to be the case."
A typically mosaic reductional trend is Regan (1909) associated the three suborders
seen in the opercular series. In Hiodon (Notop- Osteoglossoidei, Notopteroidei (including
teroidei), all Osteoglossoidei (except Panto- the Hiondontidae), and Mormyroidei within
don), and the Mormyriformes, a small to very his order Isospondyli, but he did not propose
small suboperculum is present, but this bone any formal grouping of the three.
is wanting in Pantodon and in the other mem- Garstang's (1931) rather revolutionary
bers of the Notopteroidei. classification places the families considered
Circumorbital bones are best developed in here in two divisions (Hyodontoidei and
the Osteoglossiformes (the extensive postoc- Mormyroidei) of a larger group, the Osteo-
ular bones completely cover the cheek) but glossi, in which he also placed the Elopoidei
show some reduction in the Notopteroidei and Clupeoidei. The Osteoglossi are coordi-
(especially in the postocular series of the nate with the Plectospondyli (= Ostario-
Notopteridae; those of the Hiodontidae be- physi) in Garstang's section Otophysi of the
ing somewhat more extensive). Greatest re- Malacopteri. The Otophysi were defined on
duction is found in the Mormyriformes, in the basis of the members' having, usually,
which the whole series is affected but often swim-bladder-otic connections. In view of the
differentially. The Notopteridae exhibit a very different otophysic connections of the
most peculiar development, namely, an ex- Ostariophysi, the Clupeidae, Hiodontidae,
tensive subocular shelf formed from the sec- and Notopteridae, and Megalops among the
 364 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
Elopidae, it is impossible to accept Gar- species still living in India (Sanders, 1934).
stang's Otophysi as a phyletic assemblage. The record for the Osteoglossidae is some-
Moreover, there are no substantial reasons what better. An extinct genus, Phareodus, oc-
for including the elopoid and clupeoid fishes curs in Eocene fresh-water deposits of Wyo-
in a division with the osteoglossiforms and ming, Tertiary beds in Queensland (Hills,
mormyriforms. Garstang relied almost en- 1934), and in Tertiary (?Eocene) deposits of
tirely on the presence of " . . . opposing teeth Sumatra (Sanders, 1934). From the Suma-
on hyoid and parasphenoid." This character tran beds, Sanders also identified some typ-
is developed in some elopoids but is hardly a ically osteoglossid scales as those of Sclero-
characteristic of the clupeoids, and there are pages, a genus still surviving in the Indo-Aus-
many differences between the elopoids and tralian Region. The genus Brychaetus, known
osteoglossoids which militate against close re- from fossils of Eocene age in Europe and
lationship. North Africa, is usually included in the Os-
Gregory (1933) also proposed a formal teoglossidae. As yet, the osteology of Brychae-
grouping by uniting the Notopteridae, Hio- tus does not seem sufficiently well known to
dontidae, and Mormyridae (including Gym- warrant such action. However, the genus
narchus) as a superfamily (Mormyroidea) of should probably be referred to the Osteo-
the Isospondyli. In Gregory's scheme the glossiformes, perhaps as a distinct suborder;
Mormyroidea are immediately preceded by Brychaetus may have been a marine species.
the Osteoglossoidea (comprising the Osteo- The extinct family Plethodontidae, repre-
glossidae and Pantodontidae); his only com- sented by several genera from Upper Creta-
ments on a possible relationship between the ceous deposits in Europe, North Africa, and
two superfamilies were confined to a remark North America and usually classified with the
on similarities between the pectoral girdles osteoglossoids, is, in our opinion, insuffi-
and certain aspects of syncranial architecture ciently known to warrant its inclusion in the
in Hiodon and Osteoglossum. Gregory (1933) Osteoglossomorpha.
also gave a good historical summary of views The possible relationships of at least some
on the relationship of the Osteoglossoidea and members of the extinct and supposedly
Mormyroidea. clupeoid family Ichthyodectidae (included in
Gosline (1960) suggested a formal group- the Chirocentridae by Saint-Seine, 1949, and
ing of the hiodontids, notopterids, and osteo- Bardack, 1964; see also Nybelin, 1961 and
glossids. He united them in a suborder (Os- 1964) with the Osteoglossomorpha is dis-
teoglossoidei) which forms a distinct division cussed above.
(Osteoglossi) within the Clupeiformes, all Most Osteoglossidae are large to moder-
other isospondyls being placed in the division ately large fishes with predaceous feeding
Clupei. Gosline did not give a detailed ac- habits. Exceptional are Clupisudis, which al-
count of relationships between the Osteo- though conforming to the modal size is a mi-
glossi and the Mormyriformes, but he ac- crophagous species, and Pantodon, which con-
knowledged the existence of some relation- forms to the trophic mode (it is insectivorous)
ship (1960,1961). Greenwood (1963) elevated but is small, rarely exceeding a length of
Gosline's division Osteoglossi to ordinal sta- 10 cm. All osteoglossids are robust-bodied
tus (Osteoglossiformes) and commented on fishes with a large gape; the body is either
the relative isolation of the order from other somewhat compressed (Osteoglossum and
living Isospondyli. Greenwood, like Gosline, Scleropages) or rounded to subrectangular in
acknowledged the probable interrelationship cross section, In Pantodon the pectoral fins
of the Osteoglossiformes and Mormyriformes. are greatly enlarged, and the fish skitters
The latter they thought to be derived from along the water surface.
some branch of the notopterid-hiodontid line, The Notopteroidei are also large-mouthed,
but details were not given. predaceous fishes. The Notopteridae are elon-
No fossil Mormyriformes are known and gate, laterally compressed fishes of moderate
the only fossil Notopteridae are from Lower to small size; the Hiodontidae are less com-
Tertiary beds (?Eocene) of Sumatra, the fish pressed and deeper bodied, with distinct cau-
being identified as Notopterus notopterus, a dal fins, and are of moderate size.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 365
A greater range of mouth size and form is characters shared by the two families (except
encountered among the Mormyriformes. perhaps the intrabuccal dentition) can as
Generally, the mouth is small, although some well be considered basic teleostean characters
large-mouthed species are known. The head as indicators of close phyletic relationships.
shape is extremely varied but with a trend to- In another paper Ridewood (1904b) justi-
ward elongation of the snout into a narrow, fied his rejection of Boulenger's (1898) views
decurved tube. In general, the Mormyri- of an albulid-mormyrid relationship in a
formes are somewhat compressed, dark-col- closely detailed argument (which we accept),
ored fishes of small to moderate size, and are but he did not propose any other possible re-
insectivorous. lationship except to group them loosely with
On all counts, the Osteoglossoidei repre- the notopterids and hiodontids.
sent a radiation distinct from that of the Our reasons for relating the families here
Notopteroidei and Mormyriformes, and in associated as the Osteoglossomorpha are dis-
most respects show a more primitive level of cussed above (pp. 361-363). On the basis of
organization. The mormyriform radiation is these characters we are unable to find any
undoubtedly the most specialized, but in obvious relationships outside the association.
many anatomical points members of this or- Fossil evidence may show some connections,
der show clear indications of their derivation but for the moment we assume that the Os-
from a hiodontoid-notopteroid stock. On this teoglossomorpha represent a distinct phyletic
latter line, the Hiodontidae are undoubtedly group within the teleostean level of organiza-
the most primitive members. tion.
The geographical distribution of the Osteo- Important works on the taxonomy and
glossomorpha is of considerable interest. The anatomy of the Osteoglossomorpha include
Osteoglossoidei occur in Africa (Clupisudis the following. General works are: Berg
and Pantodon), South America (Arapaima (1940), Boulenger (1898), Cockerell (1910a,
and Osteoglossum), and the Indo-Australian 1910b), Garstang (1931), Gosline (1960,
Region (Scleropages). Most members of the 1961), Gregory (1933), and Regan (1909).
Notopteroidei occur in India orAfrica (Notop- Studies on the Osteoglossoidei are: d'Auben-
teridae: Notopterus, Papyrocranus, and ton (1955; suprabranchial organ of Clupisu-
Xenomystus), but the Hiodontidae (Hiodon) dis), Daget and d'Aubenton (1957; skull of
are confined to North America. The Mor- Clupisudis), Gosline (1960; caudal fin skele-
myriformes (Mormyridae: numerous genera; ton; taxonomy), Greenwood and Thomson
and Gymnarchidae: the monotypic genus (1960; pectoral girdle of Pantodon; taxonomy
Gymnarchus) are confined to Africa, partic- of Pantodontidae and Osteoglossidae), Ny-
ularly the tropical regions. sten (1962; swim bladder and visceral anat-
As mentioned above, Ridewood (1905a) omy of Pantodon), and Ridewood (1905a;
suggested a relationship between the Osteo- syncrania of the Osteoglossidae and Panto-
glossidae and the Albulidae. There are, ad- dontidae; taxonomy of these families). Stud-
mittedly, some points of similarity in the syn- ies on the Notopteroidei are: Dehadrai (1957;
cranial architecture of the two families, espe- swim bladder in Notopterus), Gosline (1960;
cially in the well-toothed parasphenoid, pa- caudal skeleton of Hiodon; taxonomy),
latopterygoid arch, and the tongue, and in Greenwood (1963; swim bladder in the No-
the presence of subtemporal and posttempo- topteridae and Hiodontidae; skull in the
ral fossae. There are, however, many more Notopteridae; taxonomy), Omarkhan (1949b;
dissimilarities. For example, the two families sensory canals of Notopterus), Omarkhan
differ in the jaw structure, the presence or ab- (1950; chondrocranium of Notopterus),
sence of paired tendon bones, the nature of and Ridewood (1904b; syncrania of the
the caudal fin support, the ornamentation of Notopteridae and Hiodontidae; taxonomy).
the scales, and the presence or absence of a Studies on the Mormyriformes are: Daget
parasphenoid-entopterygoid articulation and and d'Aubenton (1960; chondrocranium of
of a leptocephalous stage in ontogeny. Not Mormyrus rume), Gosline (1960; caudal fin
all the differences are clearly attributable to skeleton), Omarkhan (1949a; chondrocranium
specialization in one family or the other. The of Gymnarchus), Ridewood (1904b; syncrania
 366 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 13 1
gether. In other words, some members of the
group, including the Salmonoidei, Argenti-
noidei, Galaxioidei, Esocoidei, and Stomiatoi-
dei, and some or all members of the Mycto-
phoidei, exhibit the exclusion of the maxilla
by the premaxilla from the gape, the consol-
idation of the hypural bones onto a terminal
half-centrum, the loss of an orbitosphenoid
and of a mesocoracoid, the acquisition of
photophores, and the development of an
FIG. 2. Right upper jaw bones of Dallia pec- adipose fin. Nevertheless, some myctophoids
toralis seen in front quartering view. Anterior up. have progressed further in the direction of the
Note that the toothed premaxilla extends in front Acanthopterygii than has any other salmoni-
of most of the length of the edentulous maxilla and form species. Examples of this progress are
that the lanceolate proethmoid is bound to the the elevated pectorals and advanced pelvics
symphyseal edge of the premaxilla and projects of the Aulopodidae, the acanthopterygian-
backward (downward) in approximately the orien- like jaws and hyoid and branchiostegal appa-
tation of an ascending premaxillary process. ratus of the Neoscopelidae (fig. 3), and the
berycoid-like principal caudal fin-ray for-
of the Mormyridae and Gymnarchidae), mula of numerous families, including the
Stipeti6 (1939; auditory organs of the Mor- Synodontidae and Myctophidae. Hence, the
myridae). ancestors of the living myctophoids may jus-
DIVISION III tifiably be suspected of having also contrib-
uted to the ancestry of certain major perco-
SUPERORDER PROTACANTHOPTERYGII morph types.
Several reasons lie behind our establish- This general view of percomorph ancestry
ment of the Salmoniformes as constituted be- may be further explored by a more detailed
low, and as the principal element of the Pro- consideration of the probable origin of the
tacanthopterygii. First, we are unable to find very advanced condition of the upper jaw of
a combination of anatomical characteristics Neoscopelus, a condition that is closely ap-
that satisfactorily distinguishes the Mycto- proached by the jaws of myctophids and gal-
phoidei (formerly the Myctophiformes, Sco- axiids. In Neoscopelus the premaxilla, which
peliformes, or Iniomi) from those fishes that entirely excludes the maxilla from the gape of
are usually grouped as the Salmonoidei and the open mouth, has two processes, one at its
Esocoidei. It is true, of course, that some proximal or symphyseal end, and the other on
myctophoids show various advanced features the shaft of the bone near its distal end. The
of the jaws, opercular apparatus, fins, and symphyseal process is subtriangular and ven-
caudal skeleton, but there are a great many trally concave, forming a canopy over a com-
that do not. As a result, the group has in the plex maxillary process. The symphyseal pre-
past been distinguished from the salmonoids maxillary process and the underlying maxil-
and their allies in the old Clupeiformes, sensu lary process which articulates with it have a
lato, chiefly on the superficial basis of the fact ligament that contributes to the support of
that the upper jaw is bordered by the pre- the upper jaw. The maxillary ligament ex-
maxillae alone. Such a separation cannot be tends obliquely backward and toward the
maintained, however, for this condition of the midline where it is seated on one side of the
upper jaw is to be found also in galaxiids, mesethmoid bone; the maxillary ligament is
aplochitonids, and umbrids (Dallia, see fig. crossed by the premaxillary ligament that ex-
2), among others. If one reviews the struc- tends between the premaxillary process and
tural details of the salmonoids, argentinoids, the head of the palatine. The distal premaxil-
galaxioids, esocoids, and stomiatoids, it is lary process is a thin shelf of bone on the pos-
evident that the main characteristics of the terodorsal margin of the shaft, and on which
myctophoids are merely the "advanced" fea- the maxillary shaft rests when the mouth is
tures of the first five groups collected to- nearly or completely closed. Similar processes
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 367
have been found in myctophids and ga- ascending process, at an acute angle to the
laxiids, but their symphyseal elements are not plane of that process, and from there runs
associated with a set of crossed rostral liga- postero-laterally across the dorsal surface of
ments. the body of the bone."
There is general agreement that the distal In an earlier paper, Allis (1898) had con-
premaxillary process that rests against the cluded that the ascending process of the pre-
shaft of the maxilla is the same element in all maxilla of teleosts is primarily an indepen-
fishes. The homologies of the symphyseal or dent bone, represented in the median dermal
proximal premaxillary processes (the articu- ethmoid of Amia and in the proethmoid of
lar and ascending) are equivocal, however, Esox. With this conclusion we concur, on the
and there has been no general agreement on basis of the close functional association be-
their origin, function, terminology, or occur- tween the head of the premaxilla and the
rence in fishes. The first major effort to un- dermal ethmoid. Our preparations of the skull
derstand the origin of the articular and as- of Dallia indicate that during heavy macera-
cending processes, hence to determine their tion, when the entire syncranium becomes
homologies in various groups of fishes, was more or less completely disarticulated, the
made by Allis (1909). He wrote: "[The artic- proethmoid and the premaxilla remain in
ular and ascending] . . . processes of the pre- firm contact via an extremely tough bridge of
maxillary both arise from the proximal end ligamentous tissue (fig. 2). This close associ-
of the bone, and, in the description of many ation of the two bones suggests the way in
fishes, are both included under the term as- which the ascending process may have arisen.
cending process. But one of them only is It is sometimes assumed that the develop-
properly that process, the other being a ment of an ascending process is a necessary
greatlydeveloped articular process. [The prop- concomitant of the development of upper
er ascending process is a] . . . long, thin, jaw protractility. This assumption certainly
pointed, plate-like piece of bone which lies in is unwarranted. There are many living tele-
a plane that crosses obliquely and perpendic- osts in which an ascending process of the pre-
ularly the extreme proximal end of the pre- maxilla is well developed but still attached to
maxillary. Its internal surface is presented that bone only by ligament or cartilage, for
ventro-postero-laterally and its mesial and example, in batrachoidids, lophiids, and am-
larger part rests upon and is firmly attached modytids, and in Sphyraena vulgaris accord-
by connective tissue to the corresponding ing to Allis (1909). Thiele (1963) found that
half of the grooved dorsal surface of the car- the ascending process in some blenniids is
tilaginous rostral, its mesial edge touching, separated from the premaxilla by a distinct
throughout nearly its entire length, in the line of cartilage. In each of these forms there
mid-dorsal line, the corresponding edge of its are also present on the premaxilla an artic-
fellow of the opposite side." ular process (although it may not be very
Allis also concluded that a proper ascend- distinct in blenniids) and a maxillary process,
ing process must have a ligament to the pal- but the upper jaw is significantly protractile
atine crossing an ethmomaxillary ligament. only in ammodytids.
If so, then neoscopelids, as well as cods and Indeed, the presence of ascending, artic-
their relatives, have the ascending process ular, and maxillary processes, and the asso-
much reduced and consolidated with the ad- ciation of the upper jaw with the cranium
jacent articular process, for in these forms the through a series of crossing ligaments, have
palato-premaxillary ligament is present, but made possible protractility as well as many
a distinct ascending process often is not. Al- other special functional improvements of
lis further distinguished the ascending process the upper jaw. Evidently, however, the orig-
from the articular process as follows: "The inal advantage of this complex upper-jaw
articular process of the premaxillary arises system was to provide a mechanism that
from the bone immediately lateral (distal) to combined strength with flexibility in fishes
the ascending process. It is a relatively large of primarily carnivorous habits. This view
plate of bone which lies in a plane that begins is supported by the development of a similar
at the antero-lateral edge of the base of the system among the more advanced salmoni-
 VOL. 131
368 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY

FIG. 3. Details of the skeleton of Neoscopelus macrolepidotus. A. Side view of premaxilla and maxilla,
anterior to right. B. Dorsal view of upper jaw bones, anterior up. C. Caudal skeleton, anterior to left.
D. Hyoid bar and branchiostegal rays, anterior to left.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 369
forms and among the paracanthopterygians. tainly to have achieved a myctophoid level
All these fishes are primary carnivores, with of organization. Clearly the family and its
essentially non-protractile mouths. presumed allies (aplochitonids, retropinnids,
It is of further interest to note that Allis and salangids) deserve some form of sepa-
(1909) found in Salmo a small, simple ar- rate taxonomic recognition in contrast to the
ticular process together with an " . . . inde- very much more generalized salmonids.
pendent so-called supraethmoid," and he Whereas the galaxiids have experimented
remarked on the similarities in this respect of with an advanced upper-jaw mechanism by
Salmo to Macrodon (= Hoplias) and Ery- emphasizing the articular premaxillary pro-
thrin-us of the Cypriniformes. He concluded cess, the esocoids seem to have proceeded
that the " . . . articular process of the pre- along these same lines by developing what
maxillary of teleosts would thus seem to be could be a very primitive precursor of the
as early, or even an earlier acquisition of ascending process. Esocoid caudal skeletal
that bone than the ascending process." Allis characteristics are uniformly based on a tri-
was unable, however, to find an undisputed centrum support of the hypurals. In at least
articular process in Elops or Clupea (Divi- two species the hypurals on each of the last
sion I) or in Osteoglossum (Division II). The two centra involved tend to become locked
occurrence and development of the premax- together by intervening bone and are in this
illary processes may therefore contribute im- way strongly reminiscent of the caudal struc-
portant supplementary evidence for the real- ture typical of all primitive Paracanthoptery-
ity of the Protacanthopterygii, and for the gii.
distinctness of the Salmoniformes from other Although it is not implied that esocoids may
malacopterygian fishes. have given rise to paracanthopterygians, it
In our classification of the Salmoniformes is clear that the Salmoniformes as a group
we follow Gosline (1960) in separating the contains, in mosaic occurrence, all the neces-
salmonids from the argentinoids, but we fur- sary basic specializations to have provided
ther subdivide his Salmonoidei by removing the evolutionary raw materials for all the
from that assemblage the galaxiids, aplochi- more advanced groups, including the Cteno-
tonids, retropinnids, and salangids. We place thrissiformes, of this superorder. Such, in
the latter families in a coordinate group, the fact, is the substance of our reasons for the
Galaxioidei. This separation emphasizes the establishment of both the Salmoniformes and
fact that galaxiids and their allies have under- the Protacanthopterygii.
gone a significant, if small, radiation within When considering the origin of percomorph
the Salmoniformes which is more or less com- types, Patterson (1964) evaluated the merits
parable with that of esocoids and that of myc- of myctophoids versus the Cretaceous cteno-
tophoids. Among galaxioids, esocoids, and thrissiforms as possible ancestors of the bery-
myctophoids there are some very generalized ciform fishes. He believed that the mycto-
forms and some considerably more advanced phoids could not have been directly ancestral
ones. In the Galaxiidae, one Chilean species to the beryciforms " . . . because their max-
appears to have almost as primitive a caudal illa is not toothed, while maxillary teeth per-
skeleton (fig. 4) as does any other mala- sist in some Berycoidei, and that the Cteno-
copterygian group, although its upper-jaw thrissiformes [could not have been] directly
bones are as specialized as those of the most ancestral to the group because their post-
advanced myctophids, i.e., Brachygalaxias temporal fossa is not roofed, while a partial
bullocki has its hypural bones in the caudal roof to the fossa persists in some Berycoidei."
skeleton distributed on three more or less Patterson nevertheless concluded that, though
distinct centra, but the premaxilla possesses no definite decision can be reached without
a definite articular process and entirely ex- more fossil material, the Ctenothrissiformes
cludes the maxilla from the gape. Other ga- are still in every way the most likely ances-
laxiids have the hypurals on a single, terminal tors known for the beryciform fishes.
half-centrum. In their several advanced fea- We share Patterson's confidence that the
tures, including the complete exclusion of the ctenothrissiforms or their close allies, when
maxilla from the gape, the galaxiids seem cer- these become known, will ultimately prove
 VOL. 131
370 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY

FIG. 4. Degrees of fusion in the caudal fin supports. A. Salmo gairdneri.

B. Umbra limi. C. Brachygalaxias bullocki.

to be ancestral to early acanthopterygians. sides; a scaly operculum; a caudal fin with 17

Not only are the fossil ctenothrissiforms, branched rays; and a subocular shelf on the
Ctenothrissa and especially Aulolepis, re- infraorbital bones. All these features, save the
markably similar in general appearance to maxillary teeth, and two of the branched
primitive beryciforms, as Patterson has so caudal rays, were carried forward into the
thoroughly illustrated, but one or both of generalized perciforms. These resemblances
them have in common with some or all bery- leave little doubt of the reality of the phylo-
ciforms the following characteristics: two genetic connections between the Ctenothrissi-
large supramaxillae; a premaxilla with an formes of this superorder and the Acanthop-
ascending and an articular process and an al- terygii as here defined.
veolar arm that extends below the maxilla; Another fish discussed by Patterson, the
maxillary teeth; ctenoid scales; thoracic pelvic Upper Cretaceous Sardinioides, was suffi-
fins; pectoral fins that are elevated on the ciently advanced in general structure and
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 371
sufficiently similar to the ctenothrissiforms forms are the very ones that would be ex-
to raise the question in Patterson's mind, pected to typify the ancestral paracantho-
and in ours, of the possible role this fish may pterygian. In short, in paracanthopterygians,
have played in the evolution of spiny-finned the trunk is not deep in most instances, the
groups. Of Sardinioides Patterson wrote: parietals are not well, or at all, separated, a
"Sardinioides is . . . a very generalized form, posttemporal fossa is roofed in numerous
approaching the acanthopterygians particu- species, the suboperculum is usually greatly
larly in the structure on the mouth. From the expanded, the postcleithra are reduced or ab-
analysis of acanthopterygian characters made sent, an adipose fin is present in one family,
above, three trends in acanthopterygian the alveolar extension of the premaxilla is
evolution emerged. The second of these very long and slender in many species, maxil-
trends, associated with changes in the den- lary teeth are never present, supramaxillae
tition and the development of a protrusile are absent from even the most primitive
mouth, is well advanced in Sardinioides. forms (a minute element may be present in
The first trend, associated with the use of the some ophidioids), and a subocular shelf is
pectoral fins in manoeuvering, has not pro- never present.
ceeded far. The pectoral fins have migrated In further support of a relationship be-
on to the flank, the mesocoracoid has been tween the advanced myctophoids and the
lost, and the pelvic fins have moved forwards paracanthopterygians, we have observed
(though they have not acquired an articula- the occurrence of levator maxillae superioris
tion with the cleithrum), but the trunk is not muscles in the jaw mechanism of some myc-
deep, the supraoccipital has not begun to tophids and a type of caudal skeleton with
separate the parietals, and the post-temporal dual-centrum support of the hypurals in Neo-
fossa is still roofed. The third trend ... has scopelus (fig. 3) that almost certainly must
begun with the development of feebly ctenoid have been antecedent to the extremely char-
scales. acteristic caudal skeleton of primitive Para-
"Two specialized features which might canthopterygii (see discussion below, under
exclude Sardinioides from the ancestry of the Paracanthopterygii).
acanthopterygians are the expanded suboper- Hence, we concur with Patterson that the
culum and the reduced ventral postcleithrum, Ctenothrissiformes, or their near allies as yet
and these characters link the genus with liv- undiscovered, must have been ancestral to
ing myctophoids. . . . The adipose dorsal the Acanthopterygii (as defined here), but we
fin of myctophoids [reputed to have been propose that Cretaceous myctophoids such
identified in the fossil Sardinioides] is a fea- as Sardinioides may well have produced a par-
ture which occurs in no acanthopterygians allel spiny-finned radiation which we recog-
but it is present in one family, the Percop- nize as the Paracanthopterygii.
sidae." Finally, we do not relate the ctenothrissi-
Added to the above objections to Sardini- forms to the clupeoid fishes, as has often been
oides as an ancestral type for the beryciforms done. Rather, we link them with some early
are the occurrence in Sardinioides and in group of myctophoid-like salmoniform fishes
myctophoids of an unusually long alveolar in which the supramaxillae were not reduced,
extension of the premaxilla under the maxilla the premaxilla had not excluded the maxilla
(shorter in all beryciforms), the absence of from the gape, and in which the adipose dor-
teeth on the maxilla (present in some Bery- sal fin had disappeared. The combined trends
coidei), the smallness or absence of the supra- to develop a complex upper jaw with articu-
maxillae (invariably present and large in lar and ascending processes, to elevate the
beryciforms), and the absence of a subocular pectoral fin on the flank, and to advance the
shelf (present in beryciforms). None of the pelvic fins to a position under the pectorals
above characters separating the myctophoids seem to be a peculiarly protacanthopterygian
(and Sardinioides) from the beryciforms constellation of characters that is not known
excludes them from the ancestry of the Para- in any clupeoid. Moreover, all clupeiforms,
canthopterygii. In fact, the particular char- even the most primitive (Denticeps), possess
acters by which they differ from the beryci- the highly distinctive combination of an oto-
 372 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
physic connection, opercular lateral-line ca- and preopercular canals nor do they have a
naIs, and a cranial recessus lateralis that sets clupeomorph type of ear-swim-bladder con-
these fishes entirely apart and excludes them nection. Several skull characters, such as a
from the ancestry of other teleosts. large basisphenoid, the shape and relations of
the pterotic and epiotic bones, and the large
amount of skull cartilage suggest relations
PROBLEMATICAL SUBORDERS OF THE with salmonoids. However, the caudal skele-
ORDER SALMONIFORMES ton of stomiatoids is more specialized than
SUBORDER STOMIATOIDEI: Stomiatoids are that of generalized salmonids but similar to
pelagic and bathypelagic fishes of uncertain that of the most advanced members of the
affinities. Regan (1923) considered them to Salmonoidei.
be related to clupeoids. Garstang (1931) and Among the several trends in the stomia-
Beebe and Crane (1939) suggested that they toids is the elongation of the body, even in
may have relationships with salmonoids. the relatively unspecialized Gonostomatidae.
Stomiatoids are characterized by photo- The Sternoptychidae are exceptional in hav-
phores, a large basisphenoid associated with ing a body that is extremely compressed and
the absence of lateral prootic walls in the an- deep. There is a marked tendency toward re-
terior region of the posterior myodome (see duced ossification in one or more of the ver-
fig. 5), preopercular and infraorbital lateral- tebrae immediately posterior to the skull.
line canals connected by the supraorbital ca- Even in many relatively unspecialized gono-
nal in the frontal, often a dorsal and in some stomatids and astronesthids, the cranium is
cases a ventral adipose fin, and a mesocora- attached to the vertebral column by rela-
coid in those members with well-developed tively long and extremely tough but flexible
pectoral fins. Like many clupeoids, but also ligaments. In the Astronesthidae and Melan-
salmonoids, stomiatoids tend to have bony ostomiatidae the trend toward elongate body
denticles or teeth along the ventral border of form is associated with a reduction in pectoral
the maxilla, and well-developed supramaxl- fin size. In many of the more predaceous
lae are present in the less specialized mem- groups (Gymnophotodermi and Lepidophoto-
bers. Unlike clupeoids, stomiatoids do not dermi of Parr) the gill rakers have been re-
have a close approximation of infraorbital placed by teeth, and long, canine-like, de-
sphenotic

parosphenoid
FIG. 5. Lateral view of the posterior part of the cranium of
Astronesthes niger, anterior to left.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 373
pressable teeth have developed in the jaws. Crane (1939) on the Melanostomiatidae; and
The most specialized of the many modified Grey (1960) on the Gonostomatidae. A few
jaw mechanisms that occur in this suborder is important papers on functional morphology
to be found in Malacosteus (see Gunther and have been published, for example, those of
Deckert, 1959). Deep-water forms are weakly Tchernavin (1953) on the feeding mechanism
ossified, while surface-water species are mod- of Chauliodus, Gunther and Deckert (1955)
erately bony. A tendency to lose the parietal on the functional morphology of the jaws and
bones recurs in some groups. breathing apparatus of several stomiatoids,
Stomiatoids may be divided into two or and Gunther and Deckert (1959) on the func-
perhaps three groups, the first containing the tion of the jaw apparatus in Malacosteus and
generalized, scaled Gonostomatidae and a Photostomias.
specialized offshoot of the latter family, the SUBORDER ALEPOCEPHALOIDEI: There is
Sternoptychidae. Hubbs (1953) suggested little critical anatomical information on the
uniting the Gonostomatidae and the Sternop- Alepocephalidae, and any decision concerning
tychidae. Considering the close relationship their taxonomic position must therefore be
between some gonostomatids and sternopty- considered tentative.
chids noted by Gregory and Conrad (1936) On the basis of syncranial architecture and
and accepted by Rechnitzer and Bohlke caudal fin skeleton, the Alepocephalidae can-
(1958), we suspect that a study of the mor- not be included in the Clupeomorpha as here
phology of these fishes will support Hubbs's restricted, nor, on these same characters, can
suggested merger. The remaining families they be shown to resemble the Elopomorpha.
appear to be interrelated on the basis of skull The Alepocephalidae differ mainly from
characters and in the frequent occurrence of the families of the Clupeomorpha in the ab-
a mental barbel. Furthermore, most of the sence of an otophysic connection (swim blad-
included species are without scales. Parr der not developed) or remnants of such a con-
(1927, 1930) proposed to separate the stomia- nection, in the neurocranium (no temporal or
toids into three suborders, the Heterophoto- pre-epiotic fossae, no auditory fenestrae), and
dermi [= Gonostomatoidea] containing the in the caudal fin skeleton (see Gosline, 1960).
Gonostomatidae and Sternoptychidae, the In the alepocephalids, the caudal fin skeleton
Lepidophotodermi [= Stomiatoidea], includ- has a distinct first uroneural (not fused with
ing the apparently scaled Stomiatidae and or wedged into a centrum), three distinct
the Chauliodontidae, and the Gymnophoto- centra, and a second hypural that articu-
dermi [ = Astronesthoidea] for the remaining, lates with a centrum and is not free as it is
scaleless families. Beebe and Crane (1939) in all clupeomorphs except the extremely
and Berg (1940) recognized these groups as primitive Denticipitoidei.
superfamilies. The distinction between the At least in Alepocephalus, and unlike the
Stomiatoidea and the Astronesthoidea re- Clupeomorpha, the parapophyses are fused
mains problematical. Such a separation was with the centra, and only the last haemal arch
not recognized by Regan and Trewavas is autogenous. In this genus, too, the foram-
(1930) even though outlined by Parr (1927). ina associated with the pars jugularis show an
Important works of relatively recent date advanced condition, namely, one in which
on the morphology and classification of there are no separate openings for entry of the
stomiatoids are: Regan and Trewavas (1929) head vein and the orbital artery.
on the Astronesthidae and the Chauliodonti- The most noticeable difference between the
dae; Regan and Trewavas (1930) on the Sto- Alepocephalidae and the Elopiformes lies in
miatidae, Melanostomiatidae, and Malacos- the syncranium, particularly in the palato-
teidae; Parr (1927) on the Melanostomia- quadrate arch and in the branchial skeleton
tidae; Norman (1930) on the Gonostomatidae (teeth occur only on the palatine and ecto-
and Sternoptychidae; Ege (1934, 1948) on the pterygoid in Alepocephalus) ; there are several
Stomiatidae and Chauliodontidae, respec- differences in the occipital region of the skull
tively; Beebe (1934) on the Idiacanthidae; and in the jaws (no posttemporal or subtem-
Parr (1937) on the Sternoptychidae; Schultz poral fossae, no bulla in the intercalar, pan-
(1961) on the Sternoptychidae; Beebe and etals widely separated, no separate entrances
 374 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
for the head vein and orbital artery, premax- padephori which contained the stomiatoids
illa contributing appreciably to the gape). and scopeloids, and yet another group which
The caudal fin skeleton of Alepocephalus ap- comprised only the esocoids. Gregory (1933)
proaches that of the elopiforms, but differs in actually lumped the alepocephalids and
many details. stomiatoids in a single suborder of the Iso-
Taken in sum, the characteristics of the spondyli. Obviously, much more research is
Alepocephalidae do not suggest close relation- needed before the status of the Alepocepha-
ships with any group of Division I. loidei is understood.
By contrast, we cannot find any trenchant Parr is engaged in revisional studies of the
characters to separate the Alepocephalidae alepocephalids (Parr, 1951, 1952) and has
from the Salmoniformes of Division III. The published a complete review of the Searsiidae
skull does show points of difference, but the (Parr, 1960), a group that we do not for the
degree of differentiation is considerably less present recognize as of familial rank.
than that existing between the alepocephal-
ids and Division-I fishes. The skull of Alepo-
cephalus could, indeed, be derived from a ba- SIGNIFICANCE OF ORDER GONORYNCHIFORMES
sic salmoniform type. Interpretation of the IN HISTORY OF OSTARIOPHYSAN FISHES
alepocephalid skull is made somewhat diffi- Fishes now grouped in this order have had
cult by the probability of its showing adap- a checkered taxonomic history.
tive responses to the deep-sea environment. The history of Cianos may be summarized
The caudal fin skeleton of alepocephalids is as follows: in 1775 Forskal first described a
basically like that of the "primitive" salmoni- specimen of Chanos as a Mugil; Richardson
form type, e.g., Salvelinus. The differences (1843) placed it in the genus Leuciscus.
in the caudal skeleton between the alepoce- Chanos was placed, with Gonorynchus, in a
phalids and the salmoniforms are consider- category of intermediate Malacopterygii by
ably fewer than those between the clupeo- Valenciennes (1846). Gunther (1868) consid-
morphs and elopomorphs. The caudal skele- ered Chanos to be a member of the Clupeidae,
ton of Alepocephalus shows greater affinity but this view was not accepted by Kner
with the salmoniform type than it does with (1869), Cope (1871), or Gill (1872) who con-
the other primitive types of caudal skeleton sidered Chanos as representing a distinct fam-
found in Esox, Hiodon, and Denticeps. ily separated from the Clupeidae. Near the
Because so little is known about the anat- turn of the century, Woodward (1901) placed
omy of the Alepocephalidae, we are unable Chanos in the Albulidae, a move that has
even to suggest their affinity with any par- never received support from other ichthyolo-
ticular group of the Salmoniformes. Since gists. At about the same time Ridewood's os-
there are known differences between the Ale- teological studies (1904a, 1905c) clearly dem-
pocephalidae and the Salmonoidei, we pro- onstrated that Chanos and A lbula were not
pose that, at least for the time being, the ale- related and led him to the supposition that,
pocephalids be recognized as a suborder (Ale- although Chanos was not a clupeid, it should
pocephaloidei) of the Salmoniformes. be considered a clupeoid fish. He also pointed
Regan (1929), Jordan (1923), Berg (1940), out the resemblances between Chanos and
and Norman (in his "Draft Synopsis") all re- Phractolaemus and suggested their relation-
ferred the Alepocephalidae to the suborder ship (a move amply borne out by the recent
Clupeoidei, although Berg, (1940, p. 431) work of Thys van den Audenaerde, 1961). In
commented that the Stomiatoidei were his classification of 1923, Jordan included
". . . especially near the Alepocephalidae." Chanos, together with Gonorynchus, Phrac-
Bertin and Arambourg (1958) also com- tolaemus, Kneria, and Cromeria, in the sub-
mented on alepocephalid-stomiatoid resem- order Clupeoidei, but so arranged that Cro-
blances, but placed the former in a distinct meria was distantly separated from the other
suborder (Alepocephaloidei) of their Clupei- genera. This conclusion, with regard to
formes. Earlier, Garstang (1931) treated the Chanos as a clupeoid fish, is essentially simi-
alepocephalids as a group coordinate with the lar to Jordan's earlier (1905) views, but it is
salmonoids, another group termed the Lam- interesting to note that in that book Jordan
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 375
commented on the superficial resemblances Cromeria, and Phractolaemus by many of the
between Chanos and a dace. same characters that closely unite those four
Regan (1929) followed an essentially simi- genera. All five have in common a supra-
lar arrangement, associating Chanos and Cro- branchial organ, upper and lower intermus-
meria with Kneria among the clupeoid fishes, cular bones, toothless jaws in which the pre-
but placed Gonorynchus near the Mormyri- maxilla is either a thin splint or scalelike
dae (although in 1909 he suspected a rela- bone, and a caudal skeleton with the uroneu-
tionship between Gonorynchus and Kneria). ral and lowermost hypural forming a V-
Myers (1938) suggested that Cromeria might shaped wedge around four or five smaller hy-
be a larval Kneria. In Berg's 1940 classifica- pural elements. The whole unit stems from a
tion Chanos and Kneria are included in the terminal half-centrum. Suggestive also of re-
clupeiform suborder Chanoidei, whereas lationship among the genera is the develop-
Phractolaemus and Cromeria are each given ment of a protractile upper jaw and an ex-
subordinal status but retained near the Cha- tremely delicate, loosely articulated, palato-
noidei; Gonorynchus is also given subordinal pterygoid suspension in Gonorynchus, Kneria,
rank but is placed, without comment, be- Cromeria, and Phractolaemus. The mouth is
tween the Opisthoproctoidei and Notopter- superior in Phractolaemus, inferior and below
oidei. Bertin and Arambourg (1958) united a projecting snout in Gonorynchus, Kneria,
Chanos, Kneria, and Phractolaemus within a and Cromeria. In Chanos the mouth is ter-
single clupeiform suborder (Chanoidei), but minal, non-protractile, and has a firmer and
they kept separate suborders for Cromeria more heavily ossified suspension.
and Gonorynchus. Gosline (1960) was the Other similarities among the genera have a
first author to suggest a probable relationship distinctly mosaic distribution. For example,
among all five genera, but he stressed the the infraorbital and preopercular canals of
very loose nature of their relationships by rec- the cephalic lateral-line system are similar
ognizing five superfamilies within his subor- and widely separated in all, but only in
der Gonorynchoidei. A recent paper dealing Chanos, Kneria (fig. 6), and Phractolaemus
with Chanos, Kneria, Cromeria, and Phrac- does a suprapreopercular bone develop. In
tolaemus (d'Aubenton, 1961) considered the contrast, a supratemporal commissure is pres-
problems of their relationships but reached ent in the cephalic lateral-line system (fig. 6)
no definite conclusion. Most recently Gery of all genera save Cromeria. On the other
(1964) has described a new family, genus, and hand, a fifth ossified basibranchial occurs be-
species of Kneria-like fish from African fresh tween the enlarged ceratobranchials of the
waters (family Grasseichthyidae). We had fifth gill arch (tooth-bearing lower pharyn-
previously studied this fish and concluded geal bones) only in kneriids (fig. 7), and the
that it must be included in the Kneriidae. swim bladder is divided in Chanos and
Anatomically it links Kneria with Cromeria, Kneria, simple in Cromeria, respiratory in
differing from each, however, mainly in the Phractolaemus, and absent from Gonorynchus.
absence of epipleural and cranial ribs. Be- In Chanos and Kneria the division results
cause we (Greenwood and Rosen) have not from a constriction of the swim bladder near
yet published our anatomical findings on the origin of the ductus pneumaticus as in the
Grasseichtzys, we prefer not to consider the cypriniform fishes.
genus further in this paper. The most impressive, and perhaps phylo-
Although we agree in essence with Gosline genetically the most significant, similarity
that the gonorynchoid fishes are more closely among the gonorynchiform genera is the pres-
related to one another than any of them is to ence of one or more cephalic ribs associated
other fishes, we think that Chanos, Kneria, with the specialized first three cervical verte-
Cromeria, and Phractolaemus form a close- brae. In Gonorynchus there is a single, deli-
knit and natural assemblage in which at cate, partly ligamentous rib on a posterior
least Kneria and Cromeria must be united in ledge of the exoccipital. Each of the first
a family Kneriidae. Gonorynchus is certainly three vertebrae is peculiarly modified and dif-
the most divergent member of this group, but ferent from its fellows (fig. 8). The third is es-
is nonetheless tied in with Chanos, Kneria, pecially differentiated; it possesses a deep ex-
 4-)

0.
co-4

3ri
CO O

Cie
4-)

0~
X
C

4-
C)
0
0
.0
:> L. X 4co>
L. a
_0
a 4
0 6*4
CS 0 .0
a 1

o
a- 40
- Z
C0) ._

4- 40-
s:
'4
0.
CL
L.Xa .04
0 0
44- t 4-
aS- c)
to 0 0

4, 4-4o Ca
a L. 0
4-404
.-
0. 40
._7 ._-

403-U
L Mt=
-I-

lb
C
4-43.-4
C) >
cUu
-Cv

~C)
04(

Cs4-
0

376
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 377

(-j --_,~~gill arch one

gill arch two

Q( @2 j t v gill arch three

gill arch four

gill arch five

(lower pharyngeal)

FIG. 7. Pharyngobranchial skeleton of Grasseiclthys, in dorsal

view. Note the well-ossified fifth basibranchial (in black) lying
between the lower pharyngeal bones.

cavation for the head of a large riblike bone phalic rib originates partly on the basioccipi-
bearing a posterodorsal flange. The normal tal and partly on the closely associated first
complement of pleural, epipleural, and epi- vertebra. It passes backward parallel to the
neural ribs occurs first on the fourth verte- first cephalic rib and meets a further exten-
bra. Epineurals and true ribs are not present sion of the pterotic process on the upper edge
anterior to the fourth vertebra. of the cleithrum. Halfway along its length, a
In Chanos there is an arrangement of ribs, prominence arises on the upper surface of the
muscles, and connective tissue in the occipito- second cephalic rib, from which point a broad
cervical region, which, because of their asso- strap of muscle extends forward to the exoc-
ciation with the swim bladder, suggests an in- cipital. From the posterior margin of this rib
direct otophysic connection. The first cephal- a sheet of connective passes backward to the
ic rib extends posterolaterally to the liga- anterior edge of the first, and unusually large,
mentous extension of the long pterotic pro- pleural, or true, rib on the third vertebra. The
cess; from this point it gains additional sup- second cephalic and first pleural ribs are
port from a ligament that extends anteriorly united further by a pair of modified, obliquely
to the pterotic proper. This first rib bends crossing, superficial, intercostal muscles that
posteriorly where it meets the pterotic pro- lie over the connective tissue sheet. Under-
cess and thence grades insensibly into a deli- neath this sheet a small muscle arises proxi-
cate splint which merges with the fascia of the mally on the first pleural rib and passes for-
epaxial muscles. A second, much larger, ce- ward to the occipital region. The complex sys-
378 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VO L. 131

FIG. 8. The first four vertebrae of Gonorynchus gonorynchus, showing

the peculiarly modified and movable riblike bone of the third vertebra.
Note also that the third vertebra is smaller than vertebrae in front of
and behind it, and that the fourth vertebra is the first to have a full
complement of ribs. Top: Ventral aspect. Bottom: Lateral aspect.

tem described above is especially interesting very deep, post-temporal fossa; medial to this
in view of the fact that in Chanos the first is a small fenestration bounded by supra-
pleural is firmly in contact with the upper sur- occipital, exoccipital, and epiotic bones. The
face of the anterior chamber of the swim occipitocervical complex of Chanos has, in
bladder, as in cypriniform fishes. In addi- fact, the appearance of an arrangement that
tion, as in certain Cypriniformes, there is in could precede the development of a series of
the occipital region of Chanos a large, and ossicles like those of the Weberian apparatus.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 379
Other cypriniform-like characters of Chanos finity with the Ostariophysi are the caudal fin
are the transversely subdivided swim bladder, skeleton, the occipitocervical specializations,
the form of the dorsicranium, mouth parts, and the trend toward a divided swim bladder.
and caudal skeleton, and the general dacelike In our opinion, the resemblances between the
appearance and behavior of the fish as noted two groups suggest their derivation from a
by Jordan (1905) and others. common stem, the dichotomy being very
Chanos-like occipitocervical modifications near the base. The stem itself was probably
occur in Kneria (although the two cephalic derived from some ancestral salmoniform.
ribs are fused together) in which a chambered The caudal skeleton of the Gonorynchi-
swim bladder also is present. In Cromeria, a formes, with its well-defined, V-shaped
simple, delicate cephalic rib is present, and in boundary (see above, p. 375, and Gosline,
Phractolaemus the single cephalic rib is large 1960, fig. 8) is essentially that of the chara-
and in the position of the second cephalic rib coids and cyprinoids.
of Chanos which it resembles. The occipitocervical modifications are de-
The species of Kneria, Cromeria, and scribed in detail above. The most complex
Phractolaemus occur only in the fresh waters condition of both this region and the swim
of tropical Africa. Chanos is known from salt, bladder occurs in Chanos, a genus that other-
brackish, or fresh water on both sides of the wise shows the greatest aggregation of "prim-
Pacific. Gonorynchus now occurs only in the itive" characters. Next to Chanos, the occip-
Indo-Pacific and, so far as is known, never itocervical modifications of Kneria are the
enters fresh water. Both of the latter genera, most complex, with Phractolaemus and Cro-
or at least their close allies, however, are meria showing a much simpler condition. The
known from Middle Eocene fresh-water de- picture presented by Gonorynchus is also one of
posits in North America. simplification; there is only one pair of cra-
The phyletic relationships of the Gonoryn- nial ribs and the swim bladder is wanting;
chiformes are uncertain, but they show affin- however, in this genus there is a peculiar
ity with both the Salmoniformes and the Os- modification of the pleural rib of the third
tariophysi, especially the Characoidei. De- vertebra as described above. The general im-
spite the fact that several "primitive" char- pression gained from these fishes is one of pos-
acters are preserved in some members of the sible preadaptation for the involvement of
order, there are many other characters that vertebral and neurocranial elements in the ul-
demonstrate its membership in our Division timate development of a complex otophysic
III. Among the latter may be mentioned the connection.
high degree of consolidation in the caudal fin There is some similarity between the dor-
skeleton (higher, indeed, than in many Sal- sicranial architecture of certain gonorynchi-
moniformes), the specialized jaw structure form fishes (especially Chanos, Kneria, and
(especially the loss of supramaxillae and the Phractolaemus) and that of the characoid
evolution of a protrusile mouth), and the pe- Cypriniformes. Basically the cranial lateral-
culiar occipitocervical modifications (partic- line canal system is similar in both groups, a
ularly the "cranial" ribs) often associated noteworthy resemblance being the presence
with the constriction of the swim bladder into of a well-developed supratemporal cross com-
distinct anterior and posterior chambers. The missure. The canal systems differ, however, in.
vertebral column also shows "advanced" that there is no parietal branch of the supra-
characteristics; in only one genus (Chanos) orbital canal in the Gonorynchiformes, nor is
are the parapophyses autogenous, the re- an epiphyseal branch developed on the fron-
maining genera having the parapophyses tal in these fishes, although the supraorbital
either fused with the centra (Kneria and branch in Kneria bends inward at the epiphy-
Gonorynchus) or, apparently, wanting (Phrac- seal position. These canals are usually present
tolaemus). By contrast, upper and lower in- in the ostariophysans. A suprapreoperculum
termuscular bones are present in all genera carrying the dorsalmost part of the preoper-
except Grasseichthys in which neither type of cular canal is developed in both groups.
intermuscular bone is developed. Resemblances between the Gonorynchi-
Those characteristics that suggest some af- formes and the Ostariophysi also occur in the
380 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
occipital region of the skull and in the palato- dition share with the Elopomorpha (but not
pterygoid arch. A posttemporal fossa is pres- with the Gonorynchiformes) a parietal
ent in Kneria (shallow), Phractolaemus (mod- branch of the supraorbital canal, subtem-
erately deep), and Chanos (deep), and in all poral fossae, and orbitosphenoids. In Chanos
Gonorynchiformes the supraoccipital is a there is a peculiar canal on the inner aspect
broad-based and extensive bone. In no gono- of the premaxilla. The possible homology of
rynchiform, however, is the posttemporal this canal with the apparently unique pre-
fossa divided by a transverse limb of the ep- maxillary canal in the Albulidae (see Gosline,
iotic. 1961) has been investigated. In Chanos, how-
Little guidance on the relationships of the ever, the canal houses part of an interpre-
Gonorynchiformes can be obtained from maxillary ligament.
jaw structure, because all genera except Gonorynchus is even less like an albuloid,
Chanos and Kneria show either a marked re- as was clearly demonstrated by Ridewood
duction and simplification of the jaw ele- (1905c). The persistence in certain gonoryn-
ments or, as in Phractolaemus, a highly spe- chiforms and ostariophysans of elopoid and
cialized condition. There is some resemblance albuloid characters (and, for that matter, of
between the jaw of Chanos and that of Cypri- some osteoglossomorph characters also) is yet
nus, particularly in the form of the dentary another example of the way in which pre-
and the maxilla. sumably primitive structures inherited from
As far as can be determined the jaw is pro- ancestral holostean stocks are retained in
trusible in all except Chanos, and extremely groups otherwise far removed from one
protrusile in Kneria and Gonorynchus, both of another in the teleostean radiation.
which have inferior mouths. The mouth of Boulenger (1904) considered the Phracto-
Phractolaemus is superior and highly extensi- laemidae to be most closely related to the Os-
ble, that of Cromeria subterminal. teoglossidae, but, as Thys van den Auden-
In the palatopterygoid arch of the Gono- aerde (1961) has shown, there are more differ-
rynchiformes there is only slight contact be- ences than resemblances between the two
tween the metapterygoid and quadrate, or no families, particularly in jaw structure, the
contact at all, thus recalling the fenestra be- palatopterygoid arch, and the morphology of
tween the quadrate and metapterygoid which the dorsicranium. Similarly, there are but few
is so characteristic of the Characidae (sensu shared characteristics between the gonoryn-
Weitzman, 1962) and at least some primitive chiform fishes and the Osteoglossidae or, for
Cyprinidae (e.g., Opsariichthys). that matter, the Osteoglossomorpha as a
The absence of intracranial swim-bladder whole.
diverticula, as well as the characters dis- Important papers on the anatomy and tax-
cussed above, clearly serves to exclude the onomy of the Gonorynchiformes are the
Gonorynchiformes from the clupeoid fishes classic studies of Ridewood on Chanos
(Clupeomorpha as here defined) with which (1905a), Phractolaemus (1905b), and Gono-
at least some of its members (particularly rynchus (1905c), and of Swinnnerton (1903)
Ckanos) have been associated in the past. on Cromeria. Recently there have been pa-
Chanos (see Woodward, 1901) has been pers by Thys van den Audenaerde (1961) on
linked with the Albulidae, but a detailed con- Phractolaemus, Gosline (1960) on the Kneri-
sideration of syncranial characters, branchial idae, Gonorynchidae, Chanidae, and Crome-
skeleton, and caudal fin skeletons leaves little riidae, and d'Aubenton (1961) on Cromeria.
to support such an alliance. A few details are, Giltay's (1934) anatomical studies on Kneria
it is admitted, shared by Chanos, Kneria, and are, unfortunately, marred by certain inac-
Phractolaemus on the one hand and Albula, curacies (especially in his description of the
Elops, and Megalops on the other. These in- cranium) but remain the only detailed ac-
clude posttemporal fossae, a supratemporal counts of this genus.
cross commissure in the cephalic lateral-line
system and a parasphenoid which does not SUPERORDER OSTARIOPHYSI
extend beneath the whole length of the basi- This group, consisting of some 5000 to
occipital. These characters, however, are also 6000 known species, constitutes the major-
common to the Cypriniformes, which in ad- ity of fresh-water fishes and is represented on
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 381
all major land masses except Greenland and on these continents. Siluriformes, although
Antarctica. Although the group as a whole is not confined to these continents, exhibit their
a predominantly fresh-water one, a few cat- greatest diversity there; in addition, the most
fish families (Ariidae, Aspredinidae, and primitive living catfishes are confined to
Plotosidae) contain a large percentage of southern and western South America where
marine species. Ostariophysans are of diverse they are isolated as relicts. Hence, it seems
habits and form and include characoids, likely that catfishes evolved in the South
gymnotoid "eels," carps and minnows, suck- American and African tropics. On the other
ers, hillstream fishes, and catfishes. They hand, the cyprinoids have their greatest di-
occupy a wide variety of habitats and trophic versity in southeast Asia and may have
niches. Some of the species are well-toothed, arisen there from some characoid-like ances-
predatory types, and others are omnivorous tor. The gymnotoids are confined to South
or vegetarian. Some of the most specialized America and unquestionably evolved there
forms are the toothless detritus and micro- from characoids. Regan (1922) presupposed
phagous species. The Characoidei, Cypri- a South Atlantic bridge to explain the dis-
noidei, and Siluriformes each have blind, tribution of the Ostariophysi. At times this
cavernicolous representatives. Despite the theory has been rejected by most biologists
great morphological diversity of these fishes, (e.g., Darlington, 1957) as well as by geo-
all possess a distinctive kind of otophysic con- physicists, and all evidence points against
nection, the Weberian apparatus. any Cenozoic connection between the two
Unfortunately the fossil record tells us al- continents. Piton (1938) described some fos-
most nothing about the origin and early dis- sils from the Eocene of France and considered
tribution of these fishes. Virtually all of them them characoids, thereby apparently extend-
are primary fresh-water fishes (Myers, 1938), ing the range of this group. An apparent
so that we cannot envisage a marine origin range extension into the Holarctic seemed to
for the group. The earliest fossils that can make an Afro-Asian origin of characoids more
with assurance be identified as ostariophy- feasible. However, Weitzman (1960a) dem-
sans are of Tertiary age. However, the group is onstrated that Piton's fossils are not chara-
undoubtedly much older, for all known fossil coids but that they are salmonoid or perhaps
ostariophysans are very similar to Recent esocoid relatives (Thaumaturidae) and that
species, and a very long period must have characoids are confined to Africa, South
been required for the initial differentiation of America, and Middle America northward in
the two contained orders. The relationships North America to the Rio Grande. The im-
of the Ostariophysi are discussed in connec- plications of this pattern of distribution are
tion with the Gonorynchiformes (see above). being reinvestigated in the light of recent geo-
We do not subscribe to the theory promoted physical evidence which indicates that con-
by Hoedeman (1960) that " . . . the Ostario- tinental drift may indeed have occurred (My-
physans [stand] at the very beginning of tele- ers, in press). If such earth movements took
ostean phylogeny, starting with ancestors of place and the time of the origin of this move-
the modern Siluriform series." We also be- ment can be dated, and if the characoids have
lieve that the ostariophysans are relatively always been confined to the neotropics and
primitive teleosts, but the evidence clearly in- Africa, it follows that the Cypriniformes may
dicates that characoids are more generalized have originated prior to the time when Africa
than siluriforms. Moreover, we cannot see and South America separated finally. The ap-
how a group of fishes with such a specializa- parent absence of valid pre-Cenozoic fossil
tion as the Weberian apparatus could pos- ostariophysans does not, however, preclude
sibly be ancestral to any other teleosts. their existence. The Mesozoic fresh-water
The distribution of Recent and fossil os- fossil deposits in Africa and South America
tariophysans is of great zoogeographical in- have been little investigated for fish remains,
terest and importance. Recent and fossil and, if Mesozoic ostariophysans existed, it is
characoids (Weitzman, 1960a, 1960b) are vir- in those areas that they should be found.
tually confined to Africa and to South Amer- The Weberian apparatus has been the sub-
ica, and it appears probable that most, if not ject of a great many papers, of which the
all, of their evolution must have taken place most important are cited by Alexander (1962,
382 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
1964a, and 1964b). The apparatus consists of is present in most representatives with a well-
four or more supporting anterior vertebrae developed pectoral girdle. The jaws are pro-
(pars sustentaculum) and three, and in some trusile in several groups, and in some in-
cases four, small, movable, bony parts con- stances they may have a reduced dentition or
nected to the internal ear (pars auditum). they may be completely edentulous. Pharyn-
The movable bony parts are the tripus, inter- geal teeth are highly specialized in several
calarium, scaphium, and claustrum. A cham- groups. Branchiostegal rays are few in num-
ber, the sinus cavum imparis, is joined pos- ber, usually five or fewer, but in certain cat-
teriorly and, below the foramen magnum, fishes there may be as many as 15. Scales are
with the claustrum and scaphium. This present or absent, usually cycloid, rarely
chamber is filled with fluid and is roofed by ctenoid; in a few groups scales are replaced by
inwardly directed exoccipital lamellae, and bony plates. The head is usually without
it rests on the dorsal inner surface of the scales. The swim bladder is variously modi-
basioccipital. fied, but primitively it is divided into anterior
The Weberian apparatus, present in all and posterior chambers and it usually is
Ostariophysi, is modified in many diverse physostomous. A suprapreopercular (an
and specific ways, some of which are diag- ossicle above the uppermost part of the pre-
nostic of orders, families, and other groups. opercle and partially surrounding the pre-
Superficially similar modifications of the opercular canal) is present in many species.
apparatus have been attained through con- The caudal skeleton is variously reduced but
vergence in such unrelated ostariophysans primitively and commonly with all hypurals
as the bottom-dwelling loaches and catfishes. on one centrum. There is usually an epi-
The claustrum and intercalarium tend to physeal bar between the frontals, and an
be lost from the pars auditum independently epiphyseal branch of the cephalic sensory-
in members of several families, and there is canal system of the head passes over this bar.
a distinct tendency in bottom-dwelling forms There is often a parietal branch of the ce-
for the normally two-chambered swim blad- phalic sensory-canal system.
der to be reduced and enclosed in a bony
capsule. There is also a distinct tendency for ORDER CYPRINIFORMES
centra of the Weberian vertebrae to ankylose Of the groups contained in the Cyprini-
and for additional vertebrae to be incorpo- formes, the characoids are the most primi-
rated into the apparatus. tive, and of these such forms as Brycon seem
Several other phylogenetically important among the most generalized (Weitzman,
morphological features are either present in 1962). However, Bertmar (1959, 1962) and
all Ostariophysi or else they occur mosaically. Weitzman (1964) suggest that in some ways
All ostariophysans lack an ossified basisphe- other characoids may be equally or more
noid and all appear to have an orbitosphenoid. primitive. The matter needs considerably
The saccular otolith (sagitta) is the smallest more investigation.
and either the utricular (asteriscus) or the Each of the three suborders in the Cyprini-
lagenar (lapillus) is the largest. The pelvic formes has undergone an extensive evolu-
fins are abdominal and rarely absent except tionary radiation. The basic trends and types
in a few species or as an individual anomaly. of adaptation in these groups are best de-
There is often a single spinelike ray, rarely scribed separately. Nevertheless, it is possible
two, in the dorsal fin; this fin is absent from a to describe some basic characters and trends
few species. An adipose fin is often present, in in the Cypriniformes that distinguish them
some cases with a spine at its anterior border from the Siluriformes. The Cypriniformes
(Callichthyidae and Loricariidae) and in are less specialized than the Siluriformes
others with ossified rays (Mochokidae and a despite the unusual modifications present in
few Characidae). Frequently a dorsal cranial many species. In the Cypriniformes, parietal,
fontanelle is present, bordered either by the symplectic, subopercular, and intermuscular
frontal or parietal, or by both. The skull is bones are present. A quadrate-metaptery-
usually well ossified, with little cartilage in goid fenestra is present in primitive chara-
adults even of pygmy species. A mesocoracoid coids, in most very specialized ones, but
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 383
in only a few cyprinoids (Opsariichthys atines and pterygoid bones may be toothed.
and Zacco). When fusion between vertebrae The premaxilla excludes the maxilla from the
occurs, the second and third vertebrae only gape in only one or two specialized families
are involved, In general, intervertebral (Weitzman, 1960b). An adipose fin is often
fusion is rare. The fifth vertebra is the first present. The orbital bones are usually eight
to bear a fairly normal rib, and parapophyses in number, and include an antorbital and a
usually are not fused to the centra. The supraorbital, but the series may be reduced.
body is usually scaled, rarely naked, and The caudal skeleton consists of seven hypural
heavy bony plates are never developed. The elements that are connected to the ultimate
largest otolith is usually the lagenar (as- centrum and its urostyle, and the principal
teriscus). Branchiostegal rays number three fin-ray count is almost invariably 10/9. The
to five, and vomerine teeth are absent. Weberian apparatus is relatively simple in
SUBORDER CGARACOIDEI: As noted above, most cases, seldom involving intervertebral
the Characoidei are confined to Africa and fusion, and the tripus is always attached
the Neotropical Region north to the Rio to the third centrum bya bony lamella (Weitz-
Grande in North America. There are approx- man, 1962; Alexander, 1963).
imately 23 African genera with about 150 SUBORDER GYMNOTOIDEI: The Gymno-
species, and 200 or more neotropical genera toidei are peculiar, highly specialized, elon-
with between 900 and 1000 species. gate, neotropical fishes unquestionably de-
We here recognize 16 families of characoids, rived from characoid ancestors. No extensive
all of which have been defined at the family- research on their morphology and classifi-
group level by previous authors (see Eigen- cation has been undertaken since that of
mann, 1912; Regan, 1911a; Boulenger, 1904; Regan (1911a) and that of Ellis (1913). Ana-
and Weitzman, 1954, 1960c, 1962, and 1964). tomical investigations of gymnotoid interre-
Characoids have been considered by many lationships are needed. G6ry and Vu-Tin-TuE
authors, for example, Eigenmann (1912) (1964) presented a key to the genera and
and Weitzman (1962), to belong to a single families based largely on the work of Regan,
family, the Characidae. However, the amount Ellis, and Eigenmann and Allen (1942).
of morphological divergence and the distinct- Regan's analysis is still the most complete,
ness of many groups of characoids are and it forms the basis of the discussion pre-
compelling arguments for raising several sented here.
to family rank. In recent years, interest in gymnotoids has
Characoids have undergone a great amount centered on bioelectrogenesis and object-lo-
of divergent evolution, but, despite their cation in these fishes (Couceiro and de Almei-
diversity of form, certain trends and patterns da, 1961; Lissmann, 1961), specializations
are discernible. Posttemporal fossae are pres- apparently shared with the unrelated mor-
ent and are almost always well developed. myriform fishes. In addition to having elec-
A subtemporal fossa is usually poorly devel- trogenic organs, these fishes possess several
oped, and the species that have moderately peculiar anatomical traits. They lack a rayed
developed fossae are never without jaw dorsal and pelvic fins. The caudal fin is ab-
teeth. A rhinosphenoid is often present. The sent or greatly reduced in size. The anal fin
anterior margin of the vomer is invariably has an extremely long base, and it provides
behind the ethmoid. A metapterygoid-quad- the primary means of locomotion. The basal
rate foramen is present in most forms and is pterygiophores of the anal fin are adapted
unquestionably primitive for the group. uniquely to allow a circular motion of the
Teeth are usually present on the jaws and lepidotrichia at their bases. Instead of having
are of a variety of forms (they are absent two or three sections (radials) to each pteryg-
from some families and reduced in others). Pre- iophore, the gymnotoids have only one,
ethmoids (not the proethmoids of esocoids), and in these fishes each pterygiophore has
and rostrals are absent. The pharyngeal bones a hemispherical cartilaginous head on its
and teeth are usually simple and opposing distal end which articulates directly with
but may be quite specialized in some groups the lepidotrichium. This arrangement prob-
(for example, the Chilodontidae). The pal- ably is diagnostic for the group; no char-
384 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
acoid is known to show it (Weitzman, 1962). ample, the diphyletic origin for the Gastro-
The vent is anterior, situated below the pec- myzonidae suggested by Ramaswami (1952d)
toral fins. The eyes are invariably small. The seems questionable, and we also believe that
body may be compressed laterally or rounded. the distinctions between the Gastromy-
Gymnotoids, in addition to the features zonidae and the Homalopteridae need further
mentioned above, have various other special- elucidation. The Cobitidae, Homalopteridae,
izations, such as a restricted gill opening, the and Gastromyzonidae appear related to one
reduction or absence of certain bones (in- another and may be derived from some cypri-
tercalar, palatine, pterygoid, mesocora- noid ancestor near the Cyprinidae. Rama-
coid, and suboperculum), an elongate snout, swami (1953) believed the Gastromyzonidae
and enlarged orbital bones. Despite these to be more closely related to the Cobitidae,
specializations, the relationship of the and the Homalopteridae closer to the Cypri-
gymnotoids to the characoids is evident. nidae. We await with keen interest Rama-
The Weberian apparatus is simple and essen- swami's proposed and more complete discus-
tially like that of characoids. The skull is basi- sion of the interrelationships of the cyprinoid
cally that of a characoid, especially in the families (Ramaswami, 1957, p. 302), but for
otic region. In some species the swim bladder the present we believe that the separation of
is enclosed in a bony case that differs from homalopterids and gastromyzonids as sepa-
the type found in cyprinoids. One genus rate families is not indicated. Nablant (1963)
(Gymnorhamphichthys) has a ball-and-socket has made an interesting arrangement of the
joint between the supraoccipital and the genera and subfamilies of the Cobitidae.
neural complex of the Weberian apparatus. Cyprinoids as a group exhibit adaptive
Peculiar specializations in this group are specializations not paralleled by those of
many, and a complete study may alter the any characoid. The lower pharyngeals have a
family arrangement accepted here. reduced number of teeth, and the non-opposed
SUBORDER CYPRINOIDEI: The Cyprinoidei upper pharyngeals are toothless. The former
are distributed throughout Africa, Eurasia, meet paired basiocciptial processes that often
North America, and Middle America south unite below the aorta. The tripus of the Web-
to Guatemala. There are about 250 genera erian apparatus is not attached to the cen-
and 2500 species, of which the greater number trum of the third vertebra by a bony lamella
occur in southern Asia-presumably their but is movably articulated with it. A metap-
center of origin. terygoid-quadrate fenestra is almost univer-
Despite Nichols' (1930 and 1943) opinions sally absent. Subtemporal fossae are often well
that the Catostomidae are the most primitive developed but are in some cases reduced;
of the cyprinoids, and the probable ancestors posttemporal fossae are variously developed
of cyprinids, we agree with Ramaswami but when present are not bridged by the epi-
(1957) that such is highly unlikely. The cato- otic as in the Characoidei. The preethmoid
stomid skull has several characters that and median rostral bones are variously pres-
would exclude them from the ancestry of ent or absent in accordance with the type of
other cyprinoids. The generally simpler We- jaw movements. The jaws and palatine and
berian apparatus of cyprinids lacks certain pterygoid bones are toothless, and the pre-
catostomid specializations, as described by maxilla usually excludes the maxilla from the
Nelson (1948). For us, this fact alone is gape. Orbital bones are often reduced to
enough to demonstrate the relatively primi- simple tube bones or their number is greatly
tive nature of the Cyprinidae. Moreover, increased; they are characoid-like in the more
the presence of a quadrate-metapterygoid primitive genera of the Cyprinidae. The
fenestra in two cyprinids also suggests the Weberian apparatus is often modified, the
primitive nature of these fishes (see below). most common modification being a fusion of
We look with hesitancy at some of the the second and third centra. The swim blad-
cyprinoid families proposed and discussed by der is often enclosed in a bony capsule in bot-
Ramaswami (1948, 1952a, 1952b, 1952c, tom-dwelling forms. The caudal skeleton is
1952d, 1953, 1955a, 1955b, 1957). For ex- more variable than in the Characoidei, the
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 385

FIG. 9. Medial view of left side of part of jaw suspension

of Opsariichthys bidens, anterior to right.
number of branched rays varying consider- Hemigrammocypris, Notropis, Carassius, Cy-
ably (Gosline, 1961). There is no adipose fin prinus, Hesperoleucus, Barbus (Puntius), or
except in some cobitids. Barbels are often Sarcocheilichthys.
present around the mouth. A frontoparietal Fusion of the centra of the second and third
fontanelle is present in some species, espe- vertebrae seems to be an important and spe-
cially those of the Catostomidae. cialized character of the Weberian appa-
Regan (1911a) and Ramaswami (1955a, ratus, although it may have occurred inde-
1955b) attempted to arrange the cyprinid pendently several times in cyprinids. The
genera in natural groups by sampling the character must be carefully evaluated when
osteology of one or two species of a few of groups of these fishes are defined mor-
the more than 200 nominal genera and sub- phologically. The absence of this fusion rep-
genera. Both have admitted partial failure. resents perhaps a somewhat primitive con-
From the phyletic point of view, at least dition; and its presence, a specialization.
some characters appear to have signifi- Since Barilius, as represented by B. uban-
cance at the subfamily level in cyprinids, e.g., gensis and B. pulchellus, has neither sepa-
the presence or absence of a quadrate-meta- rate second and third vertebral centra nor
pterygoid fenestra (see fig. 9). This foramen a quadrate-metapterygoid foramen, we ques-
is common in characoids (see Regan, 1911a, tion the conservative nature of Barilius pro-
and Weitzman, 1962, for figures), and from posed by Regan (1911a) and Brittan (1954).
this it may be assumed that its presence Cyprinids, for example, Esomus, with a
is primitive for cyprinids and, indeed, all laterosensory canal in the sphenotic are spe-
cyprinoids. Regan (191 la) reported it only cialized, at least in this respect (see Rama-
for Opsariichthys among the cyprinids, but swami, 1955b). Most cyprinids and characids
we found it also in the closely related Zacco. lack such a canal. This character should be
We did not find it in several samples of Ta- investigated in the genera thought to be rel-
nichthys, Rasbora, Danio (or Braci7ydanio), atives of Esomus. The ontogeny of the
Barilius, Aphyocypris (or Nicholsicypris), sphenotic and separate "dermosphenotic"
386 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 13 1

(uppermost posterior infraorbital bone) the Nematogenyinae in the Pygidiidae ap-

should be investigated thoroughly in Esomus. pear to have a simple Weberian apparatus
Perusal of the work of Ramaswami (1955a, and may be relatively primitive, at least
1955b) and Regan (1911a) reveals several for that family. We cannot accept the theory
other characters of possible significance at of Hoedeman (1954, and in litt.) that the
the subfamily, tribal, or generic level. Cau- Loricariidae are the most primitive of living
tion, however, must be used. For example, catfishes and near the ancestral "stem" of the
Ramaswami (1955b) noted the absence of Ostariophysi. It does appear that the Lori-
preethmoids in Esomus, Leuciscus, Pho- cariidae, the Callichthyidae, and possibly the
xinus, Scardinius, Aristichthys, and Notropis, Pygidiidae and a few other families may
among others. In each of the genera mentioned represent one or more distinct lines of evo-
he examined only one species, but in his lution within the catfishes. The Ariidae are
conclusions he considered those genera to be often considered relatively primitive, but
without the preethmoid. Ramaswami (loc. their lack of a mesocoracoid and fairly com-
cit.) noted the absence of the preethmoid plex Weberian apparatus casts considerable
in Notropis cornutus. Harrington (1955) doubt on this idea. Indeed, the Weberian
showed it to be present in Notropis bifrena- apparatus of the related families Bagridae,
tus, and we have found it in Notropis cocco- Ictaluridae, and Pimelodidae appears some-
genis, Notropis galacturus, and Notropis what more primitive than that of the Ariidae.
lutrensis. We have found preethmoids in Tilak (1963a, 1963b, 1963c, 1965) also has
Notropus c. cornutus, but have not re- cast serious doubt on the primitive nature of
examined Ramaswami's specimens. Possibly the Ariidae. Despite numerous recent papers
his specimens lacked preethmoids, since they (see summary at end of this section) on the
are only slightly ossified in the adult indi- morphology of catfishes, no satisfactory
viduals at hand. Although the presence or ab- arrangement of families is at present possible.
sence of this endochrondral ethmoid ossifica- The Siluriformes may be described as
tion may be of some phylogenetic signifi- Ostariophysi without a parietal, symplectic,
cance, it should be studied in closely related suboperculum, first and second pharyngo-
species and in specimens of different sizes branchials, and epipleural and epineural
before generalizations about its presence or bones. True scales are absent, and the body
absence in genera can be made. is naked or covered with bony plates. Small
spinules or prickles may be present in the
ORDER SILURIFORMES skin of some species. Short tubular ossicles
This is an unusually well-defined order of enclose the lateral line in some species.
fishes. A majority of its approximately 30 The second, third, and fourth vertebrae of the
families and perhaps 2000 species are found Weberian apparatus are fused to a single
in South America or Africa, and most are ossification called a "complex vertebra."
confined to fresh water. However, the Ariidae Additional posterior vertebral centra are
are widely distributed in tropical and sub- frequently fused to or united by exceedingly
tropical marine coastal areas, and the Ploto- tight joints with the complex vertebra. The
sidae are mainly from the Indo-Pacific. A tripus is suspended from this complex by a
few families are found in the temperate Eu- bony lamella somewhat similar to that in
rasian Region and one, Ictaluridae, is domi- characoids. A small, platelike posttemporal
nant in North America. The most primitive (when present) overlies the joint between the
of living catfish families, the Diplomystidae, pterotic and epiotic and reaches the supra-
is a relic confined to southern South America. occipital medially. The supracleithrum is
This family not only has teeth on the maxilla, complicated and articulated with the post-
but it has a relatively primitive Weberian temporal, epiotic, and often with the pterotic.
apparatus, for a siluriform (Eigenmann, Frequently, a ventral process of the supra-
1927). cleithrum (ossified transscapular ligament)
The interrelationships of the fishes of this meets the basioccipital. The lower part of
order are poorly known. It is certain that the the supracleithrum is deeply forked for the
Diplomystidae are primitive and also that reception of the upper limb of the cleithrum. A
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 387
mesocoracoid is usually present but is absent 1958, 1959), Joseph (1960), Masurekar
from some families. The first pectoral and (1962), and Tilak (1964). This list is in no
dorsal fin rays often are complicated, hard, sense complete, but, with the other works
pungent spines which frequently bear spi- cited, should serve to introduce the student
nules, serrations, pungent spinelets, or barbs. to the literature of catfish morphology and
As noted by Regan (191 lb; see also Starks, evolution.
1930, and Tilak, 1963a), the pectoral girdle
is characteristic of the group, the spine having SUPERORDER PARACANTHOPTERYGII
at its base a "friction lock" with the cor- The fishes grouped together in this super-
acoid bone. Tilak (1963a) noted a general order represent a spiny-finned radiation more
tendency in several catfish families for the or less comparable morphologically with
mesocoracoid to fuse to the coracoid, and that of the Acanthopterygii. So close are the
Starks (1930) stated that the coracoid, meso- parallels between the generalized and the
coracoid, and scapula fuse to form a complex advanced members of the Paracanthopterygii
bone. The dorsal spine also often has a char- and Acanthopterygi that the representatives
acteristic locking device, with a "locking of these two groups have not previously been
element," possibly a modified first spine, be- separated. At times, however, many of the
between the anteriorly placed basal pre- paracanthopterygians have been grouped
dorsal plate and the anterior base of the spine together as Pediculati or Jugulares. It is
itself. historically interesting that one of the orders
An adipose fin is usually present in cat- regarded by us as most primitive, the
fishes. It is in some cases preceded by a spine Batrachoidiformes, often has been treated
and rarely contains fin rays. The swim blad- as an advanced and very specialized de-
der, when well developed, is divided into an- rivative of typical acanthopterygian stock.
terior and posterior segments by a partition, The first of the toadfishes to be discovered
not a constriction as in most characoids and was described prophetically by Linnaeus as
cyprinoids. Barbels are often present and Gadus tau.
elongate, much more so than in the cyprinoids, The argument favoring the creation of this
also they frequently occur on the lower jaw. superorder is as follows: If the lophiiforms
The caudal skeleton of catfishes is variously and gobiesociforms are related to the batra-
developed, the number of principal rays choidiforms, as we have reason to believe
differing considerably. Pterygoid and palatine they may be, and if the batrachoidiforms are
bones are frequently toothed as is the vomer. the descendants of an ancient pre-acanthop-
The premaxilla and the dentary are often terygian radiation that produced also the
toothed, but the maxilla is reduced in size, percopsiforms and gadiforms, then the fishes
almost invariably toothless, and forms the here assembled evidently represent a distinct
base of a barbel. lineage that must be remote from acanthop-
Works on the morphology of catfishes terygian history.
are numerous and of widely varying quality, No acanthopterygian, not even the most
making an assessment of these and of the evo- primitive, is known to possess a caudal skel-
lution of catfishes very difficult. A great deal of eton comparable with that of the generalized
morphological work still needs to be done, paracanthopterygians, nor is it conceivable
especially in the South American families. to us how a perchlike fish could have reac-
Two of the more important works on catfish quired a pre-perciform type of cervical and
osteology and other morphology not men- caudal vertebra, and one pectoral radial (to
tioned above are the classics of Bridge and mention but three features), to produce the
Haddon (1893) and Chranilov (1929). Other characteristic batrachoidiform facies. It is
works are: McMurrich (1884), Wright (1884, everywhere evident that the trend of the
1885), Kindred (1919), Reichel (1927), David spiny-finned fishes has been toward reduction,
(1935), Shelden (1937), Nair (1938), Merri- and we view it as improbable that whole
man (1940), Eaton (1948), Bamford (1948), groups of fishes reassumed protacanthopteryg-
Krandikar and Masurekar (1954), Nawar ian-like features of the vertebrae, jaw mus-
(1954), Jayaram (1955), Srinivasachar (1957, cles, pectoral girdle, caudal skeleton, and fin
388 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VO L. 131
rays from a perchlike ancestor. One of the pterygii. The Paracanthopterygii have
earliest known perciforms, Prolates from the clearly trended toward the acanthopterygian
Lower Paleocene of the Paris Basin, has been form, and some species quite evidently have
examined by us, and we are convinced that attained that level of organization. Because
this fish had already attained a level of spe- no one group is without representatives hav-
cialization that would exclude it from the an- ing pre-acanthopterygian features, they are
cestry of anything save other perciforms. at once excluded from the acanthopterygian
It may be argued that the more primitive assemblage, and the phylogenetic integrity
Beryciformes of the Acanthopterygii could of this superorder is emphasized thereby.
have given rise to the paracanthopterygian Since paracanthopterygians have some fea-
groups. The objections to such a possibility tures that seemingly had their origin at a
are many, for example, the fact that the liv- pre-acanthopterygian level, the question may
ing and fossil beryciforms have, in common be raised whether the hypothetical "para-
with generalized perciforms but with no gadi- berycoids" and the berycoids were two dis-
form or percopsiform, the development of a tinct and quite independent radiations from
subocular shelf on the infraorbital bones, the the salmoniform level, as suggested above.
division of the hypurals into six or seven ele- The various advanced salmoniforms (such as
ments and their distribution usually on a the Myctophoidei) possess in the aggregate
terminal half-centrum, the displacement of all the necessary specializations on which to
Baudelot's ligament onto the basicranium by build each of these two basic types of spiny-
fusion of the first vertebra to the basioccipi- finned fishes (see discussion under Protacan-
tal, and the failure to develop or absence of thopterygii).
a mm. levator maxillae superioris series and The more primitive paracanthopterygian
the great development of the superficial di- orders (Percopsiformes, Gadiformes, Batra-
vision (A1) of the m. adductor mandibulae in choidiformes) are united by the following fea-
the jaw mechanism. Another obj ection to tures, present in most or all species of these
alignment of beryciforms with primitive groups, which also set them apart from the
paracanthopterygians is the demonstrably di- acanthopterygian stem: mm. levator maxil-
rect relationships between the known fossil, lae superioris present and well developed
and some living, beryciforms and the perci- (pls. 21, 22); m. adductor mandibulae super-
forms (Patterson, 1964). Patterson (per- ficialis (A1) markedly reduced or absent; in-
sonal communication) knows of no evidence fraorbital bones, when present, without a
in the fossil material that even hints at a re- subocular shelf; Baudelot's ligament from the
lationship between the Beryciformes and any posteroventral edge of supracleithrum to
group included here in the Paracanthopte- first vertebra; caudal skeleton consisting of
rygii. Moreover, Freihofer (1963) has not three hypurals only, the lowermost strutlike,
found patterns of the recurrent facial nerve the second broadly triangular and attached
in any paracanthopterygian that closely re- to a penultimate centrum, the third hypural
semble those of perciforms. Lastly, it is evi- also broadly triangular but attached to the
dent to us, as detailed below, that the Batra- terminal half-centrum.
choidiformes are more primitive than any Regan (1912a) had appreciated many of
beryciform. One might almost look upon these characteristics in his comments on a
them, together with the percopsiforms (sensu probable relationship of the batrachoidiforms
lato), as the descendants of a sort of "para- with the lophiiforms. He combined both
berycoid" radiation that gave some of its groups as suborders of an order Pediculati
main characteristics to the more specialized and he wrote: "The Batrachoidea are here in-
lophiiforms and gobiesociforms on the one cluded in the Pediculati rather than in the
hand and to the gadiforms on the other. The Percomorphi, for it can hardly be the case
essential feature of the Paracanthopterygii, that the resemblances in osteologic charac-
however, is that each group, even the most ters, especially in the structure of the pec-
advanced, contains one or more members toral arch, are not due to real affinity. In
that by current criteria are more primitive many ways the Batrachoids are more gen-
than any known member of the Acantho- eralized than the typical Pediculates, but in
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 389
some respects, notably the reduction in num- (actually meeting in the midline in some
ber of the pelvic fin-rays and the ankylosis of ophidioids), and by the discovery of a dual-
parietal and epiotic, they are more spe- centrum caudal skeleton in the brotulid
cialized. The Pediculates might certainly be Ogilbia which is precisely like that of gadoids.
regarded as highly specialized percoids, were Both gadoids and ophidioids tend to develop
it not that in the Percomorphi all the princi- complex swim bladders that are functionally
pal hypurals are attached to the last verte- related to the pleural ribs and parapophyses,
bra, whereas in the Batrachoids the upper and Freihofer (1963) has interpreted the
hypural plate is ankylosed to the last half- ophidioid pattern of the recurrent facial
centrum, and that supporting the lower half nerve as similar to that of gadiforms. The
of the fin is united to the preceding centrum, ophidioids differ from gadiforms in having
much as in the Salmopercae [ = Percopsi- the first vertebra fused to the skull, hence
formes, in part]; this seems to be a primitive in having Baudelot's ligament to the basi-
character." cranium (for example, in Genypterus blacodes,
The clingfishes resemble toadfishes in many Brotula multibarbata, and Neobythites steati-
superficial traits, and the possible relation- cus). However, in many ways they are no
ship of these two groups has been remarked more different frcm gadoids than are the
on by Starks (1905): "The families Batra- macrouroids, and we prefer to group them all
chididae and Callionymidae offer some slight together in a single order.
indications of relationship to the Gobie- Freihofer (1963) also pointed to the great
socidae, and the weight of evidence is thrown similarity of the pattern of the recurrent fa-
towards the former family by the young of cial nerve in ophidioids and zoarcids. That
some or all of them having a ventral sucking zoarcids are members of the gadiform com-
disc just behind the base of the pectorals. The plex is suggested also by the tendency of
family Batrachididae further resembles the their parietals to be enlarged and to meet in
Gobiesocidae in having the suborbital ring re- the midline (e.g., in Neozoarces, according to
duced to a small preorbital bone, only very Makushok, 1961), for Baudelot's ligament to
small parapophyses present posteriorly, no join the base of the first vertebra instead of
myodome, and a single superior pharyngeal the basicranium (e.g., in Lycodes and Macro-
present on each side. As opposing the idea of zoarces), and for the pleural ribs to be poorly
relationship the Batrachididae have five long developed. In common with many gadiforms,
actinosts, the posttemporal forms an integral the zoarcids are elongate, tailless, soft-bodied
part of the cranium, the palatine is normally fishes of the deeper waters.
joined to the pterygoid, and the mesoptery- The inclusion of the zoarcids as a suborder
goid, metapterygoid, alisphenoid, and basi- in the Gadiformes raises the question of the
branchials are present." relationships of other fishes formerly grouped
In at least some ophidioids the parietals in the Blennioidei of the Perciformes, such as
are much enlarged and meet at the midline the also elongate and soft-bodied anarhicha-
over the supraoccipital. The condition is il- dids and congrogadids. Indeed it seems to us
lustrated by Emery (1880) in Carapus. It oc- that a thorough review of the blennioids,
curs also among gadids, for example, in spe- trachinoids, uranoscopoids, and notothe-
cies of Merluccius, and it is closely ap- nioids, and perhaps even the gobioids, is
proached in a number of others, including needed before the limits of the Paracanthop-
Ra-niceps and Odontogadus (see Svetovidov, terygii and Acanthopterygii can be defined
1948). More recently it has been reported by sharply. The relationships of one or more of
Leray (1961) in the gobiesocid Gouania. those five groups to the toadfishes, codfishes,
The position of the ophidioid fishes among cusk eels, or lophiiforms were suggested or
the gadiforms was suggested by Rosen (1962) implied in the earlier anatomies or classifica-
on the basis of similarities in their jaw mech- tions of Guinther (1862), Gill (1872, 1884),
anisms, habits, and in general body form. Emery (1880), Jordan and Evermann (1898),
Their proposed relationship to the Gadi- Boulenger (1901, 1904), Regan (1903, 1912a,
formes is supported also by the trend in both 1912b), Starks (1905, 1923), Jordan (1905),
groups for the parietals to enlarge medially Gregory (1907), Goodrich (1909), and Allis
 390 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
(1909), and more recently in the work of pidids, monodactylids, and kyphosids are
Svetovidov (1948, fig. 3; 1961), Rosen among the most primitive perciforms (and in
(1962), and Freihofer (1963). some ways more primitive than serranids) be-
cause they possess, in common with beryci-
SUPERORDER ATHERINOMORPHA forms, ectopterygoid and entopterygoid teeth
The anatomical and developmental charac- and a separate foramen for the hyomandibu-
teristics of the Atheriniformes, the only con- lar trunk of the facial nerve in the wall of the
tained order of the Atherinomorpha, were pars jugularis. These perciform families bear
given by Rosen (1964). In that paper the a strong resemblance to the polymixioids
nearest relatives of the atherinomorphs were Omosoma and Berycopsis and share with Bery-
presumed to be among the ancestors of the copsis sessile ribs and scales on the bases of
Acanthopterygii. The amblyopsid fishes (Per- the median fins. Of the polymixioids, indeed
copsiformes, in part) were long associated of all beryciforms, there is a single form
with certain atheriniforms. We oppose that known, the Upper Cretaceous (Senonian)
alignment and include the Amblyopsidae in Sphenocephalus, that is generalized enough to
the most primitive surviving group of the have given rise to the serranid-percid line. On
Paracanthopterygii. the other hand various groups of specialized
SUPERORDER ACANTHOPTERYGII dinopterygoids show extensive similarities to
various specialized perciform families. For ex-
The classification of the Beryciformes pre- ample, the aipichthyids resemble carangids
sented here includes all but one group (the in having a deep trunk, long dorsal and anal
Cretaceous Dinopterygoidei of Patterson, fins with few spines and with the anterior
1964) of similar subperciform fishes that are soft rays elongate, a high supraoccipital crest
generally considered to be antecedent to the arising over the orbit, no ornamentation on
Perciformes. The dinopterygoids are omitted the head bones, upturned mouth, an elongate
from this classification simply because we supramaxilla, cycloid scales, a deeply forked
have not included exclusively fossil groups. caudal fin, and similarities in vertebral num-
However, we are fully aware that from the ber, form of cleithrum and coracoid, and in
standpoint of an enduring arrangement of the character of the interhemal and interneu-
the Acanthopterygii, the Dinopterygoidei ral radials. So close are the aipichthyids to
must be given special consideration. The rea- carangids that Aipichthys had been treated
sons were a subject of Patterson's (1964) re- as a perciform near or actually in the Caran-
cent major review of the Mesozoic acanthop- gidae, but the genus must be included with
terygians of the English Chalk. In that pa- the beryciforms because of the presence of
per, Patterson suggested that the Lampridi- an orbitosphenoid, eight branchiostegals, epi-
formes and such perciform groups as the neural bones, and 17 branched caudal rays.
Menidae, Carangidae, Acanthuroidei, Balis- The resemblances of other dinopterygoid
toidei, Chaetodontidae, and Centrarchidae families to chaetodontids, centrarchids, acan-
might be derived from the dinopterygoid thuroids, and balistoids are less impressive,
stem, and that the Serranidae, Scorpididae, but the evidence is clear. To quote Patterson:
and others might have arisen from a poly- There " . . . is evidence that the resem-
mixioid line. The Berycoidei apparently were blances between the various beryciform and
not implicated in the origin of any modern perciform groups indicate real relationship,
perciform group, nor were they particularly for if the resemblances were due to conver-
closely related to the polymixioids and dino- gence, one would expect there to be as many
pterygoids. characters in which the beryciform genera
Although some of Patterson's evidence for are more specialized than the percoid groups
a direct relationship between fossil beryci- as there are characters in which the percoid
form groups and modern perciform ones is groups are more specialized than the Beryci-
equivocal, as he himself admits, other evi- formes." Patterson goes on to remark that if
dence of his is unequivocal in its indications it is accepted that at least some of
of polyphyly among the perchlike fishes. It these different Beryciformes are near the an-
is incontestably clear, for example, that scor-
cestry of the different perciform groups which
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 391
they resemble, we have a new picture of the tion (many references of general import are
origin and evolution of the Perciformes." given in the bibliographies of Rosen, 1962 and
Patterson noted that "This type of evolu- 1964).
tion, a major group evolving polyphyletically Because of the absence at present of an
by the loss or development of characters in- over-all synthetic treatment, the subordinal
dependently and more or less contemporane- and familial classification of the Perciformes
ously in several different lines, is apparently given below is highly tentative and subject to
rather common in the history of the verte- extensive revision, although a great many of
brates. . . " and he added " . . . that the our colleagues have given us help in regard to
great majority of the changes which have the whole or parts of it. Apart from the prob-
taken place in the evolution of acanthopte- lem of the central acanthopterygians, the
rygians can be accounted for by functional composition and relationships of the Gaster-
trends. But the three main characters which osteiformes, Zeiformes, Synbranchiformes,
separate the Beryciformes from the Perci- Channiformes, and Tetraodontiformes have
formes, the presence or absence of an orbito- still to be assessed. More than one of these ex-
sphenoid and the number of pelvic and caudal tra-perciform groups is suspected of being
fin-rays, cannot be accounted for in terms of polyphyletic.
function. These three characters must have The scorpaeniform fishes represent a more
changed independently and uniformly in each or less typical example of the work that needs
of the lines of Perciformes, yet the characters still to be done. Many of these species are
are apparently nonadaptive." similar to cods in the formation of the pari-
Because of the strong indications of a poly- etals and in some respects of the cheek mus-
phyletic origin of the Perciformes raised by cles and to the toadfishes in body form and in
Patterson's work, it is evident that future na- the presence of a valved, horseshoe-shaped
tural classifications of these fishes may be swim bladder with extrinsic drumming mus-
fundamentally unlike the conventional ar- cles. Others are clearly of acanthopterygian
rangement used here. We can imagine either affinities, and those species most like the
a system in which all groups of beryciforms perciforms appear also to be the most gener-
and perciforms are united in a single order, or alized ones in the order. Although most scor-
one in which two to several new orders are es- paeniforms lack a subocular shelf on their
tablished each of which would include a series well-developed infraorbital bones, such
of basal beryciforms and the derived perci- perchlike forms as Sebastes and Sebastodes do
form types. In no case can we envision a long, have a well-developed shelf, approximately
happy future for the orders Beryciformes and 24 vertebrae, fewer than eight branchioste-
Perciformes as they now stand, a point that gals, a pelvic fin consisting of a spine and five
has already been sufficiently emphasized by soft rays, three anal spines, and principal cau-
our establishment of the Paracanthopterygii dal fin rays that never exceed 17 in number.
(see discussion above). Nevertheless, the relationships of the Scor-
Classification within the very large order paeniformes to the perciforms is not well es-
Perciformes is not in a good state at the time tablished. Freihofer (1963), for example, has
we write, although a number of workers are divided the scorpaeniforms into three well-
actively investigating either the whole group marked groups on the basis of the form and
(e.g., Freihofer, 1963, and in progress) or orientation of the recurrent facial nerve. The
parts of it (e.g., Liem, 1963, the anabantoids; distinctive pattern in scorpaenids could have
Smith, 1965, and in progress, the percoids; been derived from a widespread perciform
Haedrich, work in progress on stromateoids; pattern that typifies most species of the Ser-
B6hlke, 1960, the percoids with a disjunct ranidae. The pattern in hexagrammids, cot-
lateral line; Robins and De Sylva, work in tids, and cyclopterids could have been de-
progress on scombroids; Victor Springer, in rived from one characteristic only of Roccus
preparation, on blennies; and others). In ad- and its near allies, although the hexagram-
dition there is a great deal of past work, espe- mid-cottid-cyclopterid system of the recur-
cially on different organ systems, which con- rent facial is most like that of gobiesocids.
tains material useful for perciform classifica- The recurrent facial of the Anoplopomatidae,
392 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
however, is tentatively regarded as being Baudelot's ligament extends to the base of
closest to that of the gadoid fishes. On the ba- the first vertebra (not to the basicranium as
sis of osteological and other evidence, Quast in perciforms and some, if not all, living bery-
(1965) also concluded that the Scorpaeni- ciforms). Despite the suggestions that the
formes are complex. He wrote: "The mail- scorpaeniforms are polyphyletic, all species
cheeked fishes may be an artificial assemblage of each line that we have examined have the
containing at least three distinct evolutionary same highly distinctive type of caudal skele-
lines, scorpaenid, anoplopomatid, and hexa- ton in which two platelike hypurals are su-
grammid-cottid." The present writers have tured to the terminal half-centrum. The par-
found that at least the representatives of the ticular characters of their caudal skeleton are
hexagrammid-cottid line have a pattern of not duplicated in the species of any other
cheek muscles that seemingly could have acanthopterygian line that we have studied.
been derived as easily from the batrachoidid For the present, therefore, we retain these
as from the percoid type (pl. 23). We have fishes as a distinct order and align them ten-
also determined that various scorpaenids tatively with the Acanthopterygii.
(e.g., Scorpaena cirrhosa, S. scrofa, Pontinus
kuhlii, and Centropogon robustus) possess ' Starks (1923) claimed that the "endocleithrum"
modified epipleural ribs that extend between is not a modified rib, but the ossified Baudelot's liga-
the upper shoulder girdle and the neural arch ment. We doubt this interpretation because that bone
of the first vertebra as does the so-called en- is incorrectly oriented to be the suspensory ligament
and because a normal Baudelot's ligament is invariably
docleithrum of batrachoidids.' Moreover, in present in toadfishes in addition to the "endocleithrum."
Pontinus kuhlii and Centropogon robustus We are content to regard it as a modified epipleural rib.
 PROVISIONAL OUTLINE CLASSIFICATION
OF THE TELEOSTEAN FISHES'
DIVISION I
SUPERORDER ELOPOMORPHA
Order Elopiformes (Isospondyli in part, Clupeiformes in part)
Suborder Elopoidei
Elopidae (Elopsidae in part)
Megalopidae (Elopsidae in part)
Suborder Albuloidei
Albulidae (including Pterothrissidae, Bathythrissidae)
Order Anguilliformes (Apodes, Lyomeri, Saccopharyngiformes, Monognathiformes, Anguillomorphi)
Suborder Anguilloidei
Anguillidae
Moringuidae (including Stilbiscidae, Anguillichthyidae, Ratabouridae)
Myrocongridae
Xenocongridae (including Chlopsidae, Echelidae in part, Myridae in part, Muraenichthyidae,
Chilorhinidae)
Muraenidae (including Echidnidae, Heteromyridae)
Heterenchelyidae (Heterenchelidae)
Dysomminidae (Dysominidae)
Muraenesocidae (including Sauromuraenesocidae)
Neenchelyidae
Nettastomatidae (Nettastomidae)
Nessorhamphidae
Congridae (Congeridae, Leptocephalidae, including Heterocongridae, Colocongridae)
Ophichthidae (Ophichthyidae, including Myrophidae, Echelidae in part, Myridae in part)
Todaridae
Synaphobranchidae (including Ilyophidae)
Simenchelyidae (Simenchelidae)
Dysommidae
Derichthyidae
Macrocephenchelyidae (Macrocephenchelidae)
Serrivomeridae (including Gavialicipitidae)
Nemichthyidae (including Avocettinidae, Avocettinopsidae)
Cyemidae
Aoteidae (Aoteridae)
Suborder Saccopharyngoidei
Saccopharyngidae
Eurypharyngidae
Monognathidae
Order Notacanthiformes (Lyopomi, Heteromi, Halosauriformes)
Halosauridae

I Our attention has been called to a recent purportedly more than a few, generally superficial and insufficiently
complete list of the generic and familial names of fishes studied characters, combined with rather naive views
(Golvan, 1962). Its author does not admit to familiarity on the interpretation of palaeontological evidence. Two
with Jordan's "Classification" (1923), and the errors in other recent classifications, both in Japanese, which de-
it are exceedingly numerous. In addition, a curious and part radically from each other and from our own, were
little-known classification was published by E. Le proposed by Matsubara (1955, 1963). In these, the
Danois in 1943. Although this author's three major author follows the Stenzel system of ordinal suffixes, but
groups (orders) of teleosts are manifestly polyphyletic, his ordinal names are not incorporated in our present
his alignment of their contained suborders departs from classification. Finally, we must mention a classification
those of Regan and Berg (Berg's classification was, ap- proposed by Tretiakov (1944), which is the outline of an
parently, not seen by Le Danois). We can find little to arrangement of the orders and suborders based largely
support the reasoning behind the construction of most on the pattern of cephalic lateral-line canals. A number
of the proposed groups. The principal weakness of this of new ordinal names were proposed which are accorded
classification lies in the author's failure to consider a place in our lists of synonyms.
393
394 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VO L. 131
Lipogenyidae (Lipogenidae)
Notacanthidae
SUPERORDER CLUPEOMORPHA
Order Clupeiformes (Isospondyli in part)
Suborder Denticipitoidei
Denticipitidae (Igborichthyidae)
Suborder Clupeoidei
Clupeidae (Clupidae, including Dorosomatidae, Dorosomidae, Clupanodontidae, Dussumieridae,
Dussumieriidae, Dussumeriidae, Congothrissidae, Pristigasteridae)
Engraulidae (Engraulididae, including Stolephoridae)
Chirocentridae
DIVISION II
SUPERORDER OSTEOGLOSSOMORPHA
Order Osteoglossiformes (Isospondyli in part, Clupeiformes in part)
Suborder Osteoglossoidei
Osteoglossidae (including Arapaimidae, Clupisudidae, Heterotidae)
Pantodontidae
Suborder Notopteroidei
Hiodontidae (Hyodontidae)
Notopteridae
Order Mormyriformes (Isopondyli in part, Clupeiformes in part, Scyphophori)
Mormyridae
Gymnarchidae
DIVISION III
SUPERORDER PROTACANTHOPTERYGII
Order Salmoniformes (Isospondyli in part, Clupeiformes in part, Galaxiiformes, Haplomi, Xenomi,
Iniomi, Scopeliformes, Myctophiformes)
Suborder Salmonoidei
Salmonidae (including Coregonidae, Thymallidae)
Plecoglossidae
Osmeridae
Suborder Argentinoidei
Argentinidae (including Xenophthalmichthyidae)
Bathylagidae (including Microstomatidae, Microstomidae)
Opisthoproctidae (including Dolichopterygidae, Macropinnidae, Winteridae, Winteriidae)
Suborder Galaxioidei
Salangidae
Retropinnidae
Galaxiidae (Galaxidae, including Paragalaxiidae)
Aplochitonidae (Haplochitonidae, including Prototroctidae)
Suborder Esocoidei
Esocidae (Luciidae)
Umbridae (including Dalliidae, Novumbridae)
Suborder Stomiatoidei
Gonostomatidae (Gonostomidae, including Maurolicidae)
Sternoptychidae (Sternoptychiidae)
Astronesthidae
Melanostomiatidae
Malacosteidae
Chauliodontidae (Chauliodidae)
Stomiatidae
Idiacanthidae (including Stylophthalmidae, Stylophthalmoidae)
Suborder Alepocephaloidei
Alepocephalidae (including Platytroctidae, Platyproctidae, Searsiidae, Searsidae)
Suborder Bathylaconoidei
Bathylaconidae
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 395
Suborder Myctophoidei
Aulopodidae (Aulopidae)
Synodontidae (Sauridae, Synodidae, Bathysauridae)
Harpadontidae (Harpodontidae)
Chlorophthalmidae
Bathypteroidae (Bathypteridae, Benthosauridae)
Ipnopidae
Paralepididae (Paralepidae, including Sudidae)
Omosudidae
Alepisauridae (Alepidosauridae, Plagyodontidae)
Anotopteridae
Evermannellidae (Odontostomidae)
Scopelarchidae
Scopelosauridae (Notosudidae)
Myctophidae (Scopelidae)
Neoscopelidae
Order Cetomimiformes (Isospondyli in part, Clupeiformes in part, Stephanoberyciformes in part,
Scopeliformes in part, Ateleopiformes, Chondrobrachii, Giganturiformes)
Suborder Cetomimoidei
Cetomimidae
Barbourisiidae (Barbourisidae)
Rondeletiidae
Suborder Ateleopodoidei
Ateleopodidae (Ateleopidae, Podatelidae)
Suborder Mirapinnatoidei
Kasidoridae (Kasidoroidae)
Mirapinnidae (Mirapinnatidae)
Eutaeniophoridae (Taeniophoridae)
Suborder Giganturoidei
Giganturidae
Rosauridae (?based on young of giganturid)
Order Ctenothrissiformes
Macristiidae
Order Gonorynchiformes (Isospondyli in part, Clupeiformes in part, Chanoiformes)
Suborder Gonorynchoidei
Gonorynchidae (Gonorhynchidae)
Suborder Chanoidei
Chanidae (Chanoidae)
Kneriidae (including Cromeriidae, Grasseichthyidae)
Phractolaemidae
SUPERORDER OSTARIOPHYSI
Order Cypriniformes (Plectospondyli in part, Heterognathi, Gymnonoti, Glanencheli, Eventognathi)
Suborder Characoidei
Characidae (Characinidae, including Crenuchidae, Acestrorhynchidae, Serrasalmidae, Tetra-
gonopteridae, Creagrutidae, Glandulocaudidae)
Erythrinidae
Ctenoluciidae (Xiphostomidae, Xiphostomatidae, including Hepsetidae in part)
Hepsetidae
Cynodontidae
Lebiasinidae (including Nannostomidae)
Parodontidae
Gasteropelecidae (Gastropelecidae)
Prochilodontidae
Curimatidae (including Anodontidae)
Anostomidae
Hemiodontidae (Hemiodidae, including Bivibranchiidae)
Chilodontidae
Distichodontidae
396 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Citharinidae
Ichthyboridae (Icthyoboridae)
Suborder Gymnotoidei
Gymnotidae
Electrophoridae
Apteronotidae (including Sternarchidae, Sternopygidae)
Rhamphichthyidae
Suborder Cyprinoidei
Cyprinidae (including Gobiobotidae, Medidae)
Gyrinocheilidae
Psilorhynchidae
Catostomidae
Homalopteridae (including Gastromyzonidae, Gastromyzontidae, Lepidoglanidae)
Cobitidae (Acanthopsidae, including Adiposiidae)
Order Siluriformes (Plectospondyli in part, Cypriniformes in part, Nematognathi, Siluroidiformes)
Diplomystidae
Ictaluridae (Amiuridae, Ameiuridae)
Bagridae (including Porcidae, Mystidae)
Cranoglanididae
Siluridae
Schilbeidae
Pangasiidae
Amblycipitidae (Amblycepidae)
Amphiliidae
Akysidae
Sisoridae (Bagariidae)
Clariidae
Heteropneustidae (Saccobranchidae)
Chacidae
Olyridae
Malapteruridae (Malopteruridae, Torpedinidae, not an electric ray)
Mochokidae (Synodidae, Mochockidae, Mochochidae, Mochocidae)
Ariidae (Tachysuridae, Bagreidae, including Doiichthyidae)
Doradidae
Auchenipteridae (including Trachycorystidae)
Aspredinidae (Bunocephalidae)
Plotosidae
Pimelodidae (including Pseudopimelodidae, Callophysidae)
Ageneiosidae
Hypophthalmidae
Helogeneidae (Hologenidae)
Cetopsidae
Trichomycteridae (Pygidiidae)
Callichthyidae
Loricariidae (including Hypostomidae)
Astroblepidae (Argidae, Cyclopiidae, Cyclopidae)
SUPERORDER PARACANTEOPTERYGII
Order Percopsiformes (Microcyprini in part, Cyprinodontes in part, Cyprinodontiformes in part,
Amblyopsiformes, Salmopercae, Xenarchi, Percopsomorphi)
Suborder Amblyopsoidei
Amblyopsidae (Hypsocidae, Hypsaeidae)
Suborder Aphredoderoidei
Aphredoderidae
Suborder Percopsoidei
Percopsidae
Order Batrachoidiformes (Jugulares in part, Haplodoci, Perciformes in part, Pediculati in part)
Batrachoididae (Batrachidae)
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 397
Order Gobiesociformes (Xenopterygii, Gobiesocomorphi, Perciformes in part)
Gobiesocidae (including Diademichthyidae)
Order Lophiiformres (Pediculati in part)
Suborder Lophioidei
Lophiidae
Suborder Antennarioidei
Brachionichthyidae
Antennariidae
Chaunacidae
Ogcocephalidae (Oncocephalidae, Onchocephalidae, Malthidae, Maltheidae)
Suborder Ceratioidei
Melanocetidae
Diceratiidae (including Laevoceratiidae, Aeschynichthyidae)
Himantolophidae
Oneirodidae
Gigantactinidae
Neoceratiidae
Centrophrynidae
Ceratiidae
Caulophrynidae
Linophrynidae (including Photocorynidae, Aceratiidae)
Order Gadiformes (Anacanthini, Macruriformes, Gadomorphi, Perciformes in part)
Suborder Muraenolepoidei
Muraenolepididae (Muraenolepidae)
Suborder Gadoidei
Moridae (including Eretmophoridae, Tripterophycidae)
Bregmacerotidae
Gadidae (including Gaidropsaridae, Ranicipitidae)
Merlucciidae
Suborder Ophidioidei
Ophidiidae (including Brotulidae, Brotulophidae, Aphyonidae)
Carapidae (Fierasferidae, Disparichthyidae)
Pyramodontidae
Suborder Zoarcoidei
Zoarcidae (including Lycodidae, Lycodapodidae, Derepodichthyidae)
Suborder Macrouroidei
Macrouridae (Macruridae, Macrouroididae, Coryphaenoididae, including Lyconidae)
SUPERORDER ATHERINOMORPHA
Order Atheriniformes (Synentognathi, Beloniformes, Gambusiformes, Microcyprini in part, Cyprino-
dontiformes in part, Percesoces in part, Mugiliformes in part, Mugilomorphi in part, Phallo-
stethiformes, Perciformes in part)
Suborder Exocoetoidei
Exocoetidae (including Hemiramphidae, Hemirhamphidae, Oxyporhamphidae, Evolantiidae)
Belonidae (Esocidae, Esocesidae as of Rafinesque, including Tylosuridae, Petalichthyidae)
Scomberesocidae (Scombresocidae)
Suborder Cyprinodontoidei
Oryziatidae
Adrianichthyidae
Horaichthyidae
Cyprinodontidae (including Fundulidae, Orestiidae, Empetrichthyidae)
Goodeidae (including Characodontidae)
Anablepidae (Anablepsidae)
Jenynsiidae (Fitzroyiidae, Fitzroyidae)
Poeciliidae (including Tomeuridae)
Suborder Atherinoidei
Melanotaeniidae (including Zanteclidae, Neoatherinidae)
Atherinidae (including Bedotiidae, Pseudomugilidae)
 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
398
Isonidae
Neostethidae
Phallostethidae
SUPERORDER ACANTHOPTERYGII in part)
Order Beryciformes (Xenoberyces, Berycomorphi, Berycoidei in part, Stephanoberyciformes
Suborder Stephanoberycoidei
Stephanoberycidae
Melamphaeidae (Melamphaidae, Melamphasidae)
Gibberichthyidae
Suborder Polymixioidei
Polymixiidae
Suborder Berycoidei
Diretmidae
Trachichthyidae (including Hoplopterygidae, Sorosichthyidae)
Korsogasteridae
Anoplogasteridae (Caulolepidae)
Berycidae
Monocentridae
Anomalopidae
Holocentridae (Holocenthridae)
Order Zeiformes (Zeomorphi, Zeoidei)
Parazenidae
Macrurocyttidae (Zeniontidae)
Zeidae (including Cyttidae, Cyttopsidae, Zenidae)
Grammicolepididae (Grammicolepidae)
Oreosomatidae
Caproidae (Caprophonidae, including Antigoniidae)
Order Lampridiformes (Selenichthyes, Allotriognathi)
Suborder Lampridoidei
Lampridae (Lamprididae)
Suborder Veliferoidei
Veliferidae
Suborder Trachipteroidei
Lophotidae
Trachipteridae (Trachypteridae)
Regalecidae
Suborder Stylephoroidei
Stylephoridae (Stylophoridae) part,
Order Gasterosteiformes (Lophobranchii, Thoracostei, Aulostomi, Solenichthyes, Scleroparei in
Syngnathiformnes, Aulostomiformes, Rhamphosiformes)
Suborder Gasterosteoidei
Gasterosteidae (Sclerogenidae in part)
Aulorhynchidae
Indostomidae
Suborder Aulostomoidei
Aulostomidae
Fistulariidae (Fistularidae)
Macrorhamphosidae (Macroramphosidae, Rhamphosidae)
Centriscidae (Amphisilidae)
Suborder Syngnathoidei
Solenostomidae (Solenostomatidae, Solenostomatichthyidae, including Solenichthyidae)
Syngnathidae (including Hippocampidae, Siphostomidae)
Order Channiformes (Labyrinthici in part, Ophiocephaliformes)
Channidae (Ophicephalidae, Ophiocephalidae, including Parophiocephalidae)
Order Synbranchiformes (Symbranchia, Symbranchii, Symbranchiformes, Alabiformes)
Suborder Alabetoidei
Alabetidae (Alabidae, including Cheilobranchidae, Chilobranchidae)
Suborder Synbranchoidei
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 399
Synbranchidae (Symbranchidae, including Flutidae, Monopteridae)
Amphipnoidae
Order Scorpaeniformes (Cataphracti in part, Scleroparei in part, Pareioplitae, Loricati, Sclerogeni,
Cottomorphi, Perciformes in part)
Suborder Scorpaenoidei
Scorpaenidae (Sclerogenidae in part, including Tetrarogidae)
Triglidae (Sclerogenidae in part, including Peristediidae, Peristediontidae)
Caracanthidae
Aploactinidae (Aploactidae, ?including B'athyaploactidae)
Synancejidae (Synanceidae)
Pataecidae (including Gnathanacanthidae)
Suborder Hexagrammoidei
Hexagrammidae (including Ophiodontidae, Oxylebiidae, Chiridae)
Anoplopomatidae (Anoplopomidae, including Erilepidae)
Zaniolepididae (Zaniolepidae)
Suborder Platycephaloidei
Platycephalidae (including Bembradidae, Bembridae, Parabembridae)
Suborder Hoplichthyoidei
Hoplichthyidae (Oplichthyidae)
Suborder Congiopodoidei
Congiopodidae (Agriopidae)
Suborder Cottoidei
Icelidae (including Ereuniidae, Marukawichthyidae)
Cottidae (Sclerogenidae in part, including Jordaniidae, Blepsiidae, Blepisiidae, Scorpaenichthyi-
dae, Ascelichthyidae, Synchiridae, Rhamphocottidae, Hemitripteridae, Neophrynichthyidae)
Cottocomephoridae (including Abyssocottidae)
Comephoridae
Normanichthyidae
Cottunculidae
Psychrolutidae
Agonidae (including Aspidophoroididae, Aspidophoridae)
Cyclopteridae (including Liparopidae, Liparidae, Lipariidae, Eutelichthyidae, Rhodichthyidae,
Cyclogasteridae)
Order Dactylopteriformes (in part Craniomi, Scleroparei, Cataphracti, and Perciformes)
Dactylopteridae (Cephalacanthidae)
Order Pegasiformes (Hypostomides, Perciformes in part)
Pegasidae
Order Perciformes (Percomorphi in part, Holconoti, Labyrinthici in part, Chromides, Pharyngognathi,
Gobioidea, Jugulares in part, Malacichthyes, Icosteiformes, Percesoces in part, Mugiliformes
in part, Polynemiformes, Rhegnopteri, Bathyclupeiformes, Xenoberyces in part, Berycoidei in
part, Beryciformes in part, Thunniformes, Plecostei, Scombriformes, Echeneiformes, Disco-
cephali, Mastacembeliformes, Opisthomi, Chaudhuriiformes, Anabantiformes, Blenniiformes,
Trachiniformes, Gobiiformes, Carangiformes, Acanthuriformes, Squamipenes, Embiotoco-
morphi, Gadopseiformes, Coryphaeniformes, Amphiprioniformes)
Suborder Percoidei
Centropomidae (Oxylabracidae, including Latidae, Chandidae, Ambassidae, Ambassiidae)
Serranidae (including Percichthyidae, Chromileptidae, Perciliidae, Moronidae, Oligoridae, Mac-
cullochellidae, Macquariidae, Niphonidae, Plectroplitidae, Epinephelidae, Cephalopholidae,
Bostockiidae, Diploprionidae, Rainfordiidae, Hypoplectrodidae, Plectropomidae, Anthiidae,
Ostracoberycidae, Paracentropristidae)
Grammistidae (including Rypticidae)
Pseudochromidae
Pseudogrammidae (including Rhegmatidae)
Grammidae (including Stigmatonotidae)
Plesiopidae (including Pharopterycidae)
Pseudoplesiopidae
Anisochromidae
Acanthoclinidae
400 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
Glaucosomidae
Theraponidae (Teraponidae, Terapontidae, Tesapontidae)
Banjosidae
Kuhliidae (Duleidae, including Nannatherinidae, Nannopercidae)
Gregoryinidae (?based on young of cheilodactylid)
Centrarchidae (including Elassomatidae, Elassomidae, Cristidae, Grystidae, Micropteridae,
Eucentrarchidae)
Priacanthidae
Apogonidae (including Ostorhinchidae, Gymnapogonidae, Apogonichthyidae, Henicichthyidae,
Henichthyidae, Dinolestidae, Cheilodipteridae, "Amiidae," Epigonidae)
Acropomatidae (Acropomidae)
Percidae (including Etheostomatidae, Etheostomidae)
Sillaginidae
Branchiostegidae (Latilidae, including Malacanthidae)
Labracoglossidae
Lactariidae
Pomatomidae (including Scombropidae, Scombropsidae)
Rachycentridae (Rhachycentridae, Elacatidae)
Echeneidae (Echeneididae)
Carangidae (including Seriolidae, Nematistiidae, Juvenellidae)
Coryphaenidae
Formionidae (Formiidae, Apolectidae)
Menidae
Leiognathidae (Liognathidae, including Equulidae)
Bramidae (including Steinegeriidae, Trachyberycidae, Pteraclididae, Pteraclidae, Lepidotidae,
Lepodidae)
Caristiidae (Elephenoridae)
Arripidae (Arripididae)
Emmelichthyidae (including Erythrichthyidae, Erythroclidae, Dipterygonotidae, Inermiidae,
Maenidae, Spicaridae, Merolepidae, Centracanthidae, Centracantidae)
Lutjanidae (Lutianidae, Luthianidae, including Hoplopagridae, Etelidae, Verilidae, Aphareidae,
Caesionidae, Caesiodidae)
Nemipteridae (including Scolopsidae)
Lobotidae
Gerridae (including Eucinostomidae, Xystaemidae)
Pomadasyidae (Pomadasidae, Haemulidae, Haemulonidae, including Gaterinidae, Pristipomidae,
Pristipomatidae, Plectorhynchidae, Xenichthyidae)
Lethrinidae (including Monotaxidae, Neolethrinidae)
Pentapodidae
Sparidae (including Denticidae, Pimelepteridae, Paradicichthyidae, Paradichthyidae)
Sciaenidae (including Otolithidae)
Mullidae
Monodactylidae (Psettidae)
Pempheridae (including Leptobramidae)
Bathyclupeidae
Toxotidae
Coracinidae (Dichistiidae)
Kyphosidae (Cyphosidae, including Scorpididae, Scorpidae, Parascorpidae, Girellidae)
Ephippidae (including Chaetodipteridae, Platacidae, Ilarchidae, Drepanidae, Drepanichthyi-
dae)
Scatophagidae (including Prenidae)
Rhinoprenidae
Chaetodontidae (including Pomacanthidae)
Enoplosidae
Pentacerotidae (Histiopteridae)
Nandidae (including Polycentridae, Pristolepidae)
Oplegnathidae (Hoplegnathidae)
Embiotocidae (Ambiotocidae, including Ditremidae, Hysterocarpidae, Holconotidae)
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 401
Cichlidae (Chromidae, not a pomacentrid)
Pomacentridae (Abudefdufidae, Glyphiodontidae, Ctenolabridae, including Amphiprionidae,
Chromidae, Premnidae)
Gadopsidae
Cirrhitidae
Chironemidae
Aplodactylidae (Haplodactylidae)
Cheilodactylidae
Latridae (Latrididae)
Owstoniidae
Cepolidae
Suborder Mugiloidei
Mugilidae
Suborder Sphyraenoidei
Sphyraenidae
Suborder Polynemoidei
Polynemidae
Suborder Labroidei
Labridae (Cyclolabridae, including Coridae, Neolabridae, Bodianidae, Harpidae)
Odacidae (including Siphonognathidae, Neodaciidae, Neoodacidae)
Scaridae (Callyodontidae, including Sparisomidae, Scarichthyidae)
Suborder Trachinoidei
Trichodontidae
Opisthognathidae
Bathymasteridae
Mugiloididae (including Pinguipedidae, Parapercidae, Parapercichthyidae)
Cheimarrhichthyidae (Chimarrichthyidae, not a catfish)
Trachinidae (including Callipterygidae)
Percophididae (Percophidae, including Bembropsidae, Bembropidae, Pteropsaridae, Hemero-
coetidae)
Trichonotidae
Creediidae
Limnichthyidae (Limnichthidae)
Oxudercidae
Leptoscopidae
Dactyloscopidae
Uranoscopidae (including Astroscopidae, Pleuroscopidae)
Champsodontidae
Chiasmodontidae
Suborder Notothenioidei
Bovichthyidae (Bovichtidae, Bovictidae, Pseudaphritidae)
Nototheniidae (including Harpagiferidae, Gelididae)
Bathydraconidae
Channicthyidae (Channichthyidae, Chaenichthyidae)
Suborder Blennioidei
Blenniidae (including Runulidae, Salariidae, Atopoclinidae, Xiphasiidae, Nemophididae)
Anarhichadidae (Anarrhichadidae, including Anarrhichthyidae)
Xenocephalidae
Congrogadidae (including Halidesmidae, Haliophidae)
Notograptidae (including Stichariidae)
Peronedysidae (Peronedyidae)
Ophiclinidae (Ophioclinidae)
Tripterygiidae (Tripterygiontidae)
Clinidae (including Paraclinidae, Xenopoclinidae)
Chaenopsidae (including Emblemariidae)
Stichaeidae (including Lumpenidae, Xiphisteridae, Xiphidiontidae, Chirolophidae, Cebedich-
thyidae, Cryptacanthodidae, Cryptacanthidae)
Ptilichthyidae
402 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
Pholididae (Pholidae, including Opisthocentridae)
Scytalinidae (Scytaliscidae)
Zaproridae
Suborder Icosteoidei
Icosteidae (Acrotidae)
Suborder Schindlerioidei
Schindleriidae
Suborder Ammodytoidei
Ammodytidae (including Bleekeridae, Bleekeriidae)
Hypoptychidae
Suborder Callionymoidei
Callionymidae (including Draconettidae)
Suborder Gobioidei
Gobiidae (including Eleotridae, Milyeringidae, Doliichthyidae, Benthophilidae, Gobiomoridae,
Sicydiaphiidae, Apocrypteidae, Periophthalmidae)
Rhyacichthyidae (Platypteridae)
Kraemeriidae (Psammichthyidae)
Gobioididae (including Amblyopidae, Taenioidae, Taenioididae)
Trypauchenidae
Microdesmidae (Cerdalidae, including Pholidichthyidae, Gunnellichthyidae, Paragobioididae)
Suborder Kurtoidei
Kurtidae
Suborder Acanthuroidei
Acanthuridae (Hepatidae, Acronuridae, Harpuridae, Teuthidae, Teuthididae in part, including
Zanclidae, Nasidae)
Siganidae (Theutyidae, Teuthididae in part, Amphacanthidae)
Suborder Scombroidei
Gempylidae (Acinaceidae, including Lemnisomidae, Ruvettidae)
Trichiuridae (including Lepidopidae)
Scombridae (Scomberidae, including Cybiidae, Cibiidae, Thunnidae, Katsuwonidae, Scombero-
moridae, Sardidae, Acanthocybiidae, Gasterochismidae)
Xiphiidae
Luvaridae (Dianidae)
Istiophoridae (Histiophoridae, including Tetrapturidae, Makairidae)
Suborder Stromateoidei
Centrolophidae (including Icichthyidae)
Nomeidae (Psenidae)
Stromateidae (Pampidae)
Tetragonuridae
Suborder Anabantoidei
Anabantidae
Belontiidae (Belontidae, Polyacanthidae)
Helostomatidae (Helostomidae)
Osphronemidae (Osphromenidae, Labyrinthicidae)
Suborder Luciocephaloidei
Luciocephalidae
Suborder Mastacembeloidei
Mastacembelidae (Rhynchobdellidae)
Chaudhuriidae
Order Pleuronectiformes (Heterosomata)
Suborder Psettodoidei
Psettodidae
Suborder Pleuronectoidei
Citharidae
Scophthalmiidae
Bothidae (including Paralichthyidae)
Pleuronectidae (Planidae, including Hippoglossidae, Samaridae, Paralichthodidae, Rhombo-
soleidae)
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 403
Suborder Soleoidei
Soleidae (including Achiridae, Trinectidae, Synapturidae)
Cynoglossidae
Order Tetraodontiformes (Plectognathi, Diodontomorphi)
Suborder Balistoidei
Triacanthidae (including Triacanthodidae)
Balistidae (including Monacanthidae, Aluteridae, Aleuteridae, Psilocephalidae, Anacanthidae,
not a sting ray)
Ostraciontidae (Ostraciidae, including Aracanidae)
Suborder Tetraodontoidei
Tetraodontidae (Tetrodontidae, Gymnodontidae, including Lagocephalidae, Chonarhinidae,
Xenopteridae, Canthigasteridae, Tropidichthyidae, Ovoididae, Colomesidae, Sphoeroididae)
Triodontidae
Diodontidae
Molidae (Orthagoriscidae, Triuridae)
404 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Elopidoe
M uraenidae

H.terenchelyidoe
Megolopidoe

Albulidoe Dysomminidae

Muroenesocida.
Anguillidao

&~~ ~ -_~~~~~~~~~~~~~~~~~~~~~~~~~
~~~~
I_,
Morngud_
Neonchelyidoe

Xenoc ongridae
Nottastomatidae
CHART 1.
966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 405

Cong ridoe
Nemichthyida*

Ophichthidae Cyemidae

Synaphobranchidae
Saccopharyngidoe

Simenchelyidoe
Eurypharyngidas

Dys ommides
Monognathidae

Derichthyida.
Halosauridae

Macrocephenchylidue Lipogenyidae

Serrivomaridoe Notacanf hidae

CHART 2.
406 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VO L. 131

D.nticipitidae
Pontodontidae

Clup.idoe
Hiodontidae

Engrouli doe Notopteridae

Chirocentridoc Mormyridoe

OSt0oglossidas Gymnarchidoe
CHART 3.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 407

Salangido.
Salmonidae

Retropi nnidae

Plecoglossidae

I-)

GoIoK.iicIoke
Osmeridae

Aplochi tonida*
Argentinidoe

Bathylogidae

Opisthoproctidae Umbridao
CH9ART 4.
408 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Gonostoma fidce
Alepoc*phalidoo

Bothylaconidae
,ternoptychido.

Sfernopfychidae

Asfronesthidae
Aulopid idae

Melanostomiotidae
Synod on t ido.

Malacost.idce

H
orpadontidae
Chouliodontidae

Chlorophthalmidea
Stomiatidae

CHART 5.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 409

Ipnopidoe

Neoscopelidae

Parolepid ide

Cetofnimidoe
Omosudidae

*00

Bur bour i si.d .

Alep souridlas

Anotopterido.
Rondeletiidoe

I 2-)-
E vermonnell idas
Ateleopodida*

Scopelorchidao

Miropinnidoe

Scope 1osaurid ao

Myctophidoe Eu to eniophorida
CHaART 6.
410 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Giganturidae Characidae

Rosourida.

ErythrinidQe

Ctenoluciidae
Macristiidte

Gonorynchidae
Hepsetidae

Cynodontida.
Chanidae

Iner;;dae

K(n.riidae

Phroctol aemidoa
CHART 7.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 411

Gasteropelecidas

Distichodontidae

Prochi lodont i doe

Cithori ni dae
Curimvftidae

I chthyboridee

A nost omi dae

Gymnotido.

Hemi odontidae El* ct roplhori doe

Apteronotidae

Rhomphichthyidoe
Chilodontidae

CHART 8.
412 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Cyprinid a.
Diplomnysfidca

Gyrinochoilidao
Ictoluridae

Psilorhyn chido.
Bagridae

Cafoltos ida*
Cronoglanid ida.

Homalopferidae
Si lurida.

Cobitida* Schilbeido.
CHART 9.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 413

Chocid ae

Pongasiid a

Olyri doe
Amblycipitidoe

_alopteruridoe
Motopterurido.

Amphilii doe

Akysidae

Mochokidao

Sisori dae

Ar ii doe

Clariidae

Hetcropneustida e Dora di doe

CHART 10.
414 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Auchen;pferidae H-cIlogenei doe

Aspredinidoe
Cetop3idae

Plotosi doe
Tn cl=o mycterid)
Tri cho mycteridoe

Pimelodidoe
Callichthyidoe

Agen.;osidoe

Loricor idoe

1--

Hypophthcalmidoe Astroblepidoe
CHART 11.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 415

Amblyopsid_ s

Amblyopsida.

Ant.nn riidaed

Aph redoderida e
Chaunacida.

Percopsidae

Ogcocepholidue

Batrachoidid@e

Melanocetidoe
Gobiesocidoo

Diceratii do.
Lophii doc

CHART 12.
416 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Oneirodidae

Ceratiidae

Giga ntactinid@e

Caulophrynidae

Neoceratii dae

Linophrynidae

C ntrophrynidcae
CHART 13.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 417

M uroenolepididae Ophidiidoe

Moridoe Carapidoe

Bregmocerotidaoe Pyromodontidac

Godidae Zoo rc ida e

Merlu ccii doe Macrouridoe

CHART 14.
418 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL.
131

~~~~~~~J X
Exocoet idoe
Anoblepidae

Exoccefidae Jenynsaidao

Belonida.Q'eui

Scomnberesacida. Poeciliida.

M.Ion.t0*n;; do.

Adrionichthyida. Atberinidae

Horaichthyidae
1so doa
n

Cypr ino dontidoe Neootethidae

Goode idae Pholloatethido.

CHART 15.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 419

Stephanoberycidae

Trachichthyido.

Macrurocytti doe

CHART 16.
420 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

V Lampricide

Veliferidace

Grammicolepididca
Loph o tid a*

Oreosomcauidae

IN /-
ju V.

Stylephoridae

CHART 17.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 421

Gosterosteidae

Aulorhynchidce
Syngnothidoe

I ndostomidoe

Chonnidoe
Au lost omidoc

Alobetidae
Fistulor iidae

M ocrorhamphosidoe

Centriscidoe

Solenostomidoe
CHART 18.
422 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Scorpoenidot
Potaccidoe

Trig lid a.
H.ex agra mmi d ac

Carocanthidae

Anoplopomat idae

Aplo*c ti ni does

Zoniolepididae
Synancejidoe
CHART 19.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 423

Plctycephalidae

Cottunculidoe

Hoplichthyidae

Psychrol utidao

EAgonide

Cyclopteridoe

Icelidoe

Nmrnonichfhyido.
CHART 20.
424 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VO L. 131

P seudochromidae

Pseudogrommida*

Grommi dcxe

Kuhliidae

CrART 21.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 425

Bronchi3ostgidoe

Labro cog lossidoi

Lactorii dae

Pomotomido.

Rachycentrida e

EchIneid@e

Sila ginidae Coryphoenidat

CHART 22.
426 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

EmmeUlichthyidae

Lufjainida.

CHART 23.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 427

Kyphosidue

Mull idme

Ephippida.

CHART 24.
428 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Nondidae

Chaetodontid a

E mbiotoci doe

C ichI ida *

E noplosidae

Pomacentri doe

Godopsidae
CHART 25.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 429

Cepoli de

Cirrhiti do.

M ugiI da

3
7

Sphyra.ni dae

Aplodoctylidae
1Jc52

Labridae

C h*ildoctyli dae
Odacida,

Latridae \4 ~~~~Scaridat

O Wat on doe Trichodon i dae

C1HART 26.
430 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

L imnich
thyida
IVVA

,m,

Bothymasteridae L .ptoscopida.

Mugiloididae ayfosc opi dae

DaC

Ur a nos cop da.

Ch eimarr hicht hyi dao

C hampsodontidoe
_ Ab~ /
Percophididas

Bovichihyidae

Trichonot id a

Creed ii d ae Notothenii doe

CHART 27.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 431

{lothydraconidae

Channicththyi doe

C ho enopsidoe

Blenniidao

Stichoeidoe

Anorhichodi doe

Ptilichthyidae
Congrogodidoc

Pholididoe
fZZIZZ
Scytolinidoe
Notogra pfidoa

Z oproridoe

Peronedysidoe
Ic ost;idoe

Ophiocli ni dae Schin d er ;i doe

CHART 28.
432 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Ammodytidc-
M i crc des mid ae

Hypoptychidai

Acanthurid a-

Siganidae

Rhyacicht hyidao

Gempyli doa
K roemerii dca

GobioididaX
Trichiurida.

Try pauchenidee Scombridoa

CHART 29.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 433

Xiphiidoe
Anabontidas

Luvcridoe

B el ontiidae

Istiophor idae

Helostomoatidae
Centrolophidoe

Nomoidas

Osphron.mida.

Stromatei doe

Totragonuridas Luciocepholi doe

CHART 30.
434 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131

Mlatcenmbelidae

Bothido e

Choud huriidoe

Pleuronectidao

Psettodidoe

Cithori dae
Solsi doe

Cynoglo3sida e
Scophtho Imidoe
CHART 31.
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 435

Ostraciontidoa

Tr ia cant Kid ae
Ostrociontidae

Tetroodonfido.
Triaconthidoe

Balistidoe

Triodontidoe

Diodontidos
Balistidoe

0
Molidoe
B s d

Balistida*e
CHART 32.
 BIBLIOGRAPHY
ALEXANDER, R. McN. Gegenbaurs Morph. Jahrb., Leipzig,
1962. The structure of the Weberian appara- vol. 39, pp. 526-644.
tus in the Cyprini. Proc. Zool. Soc. Lon- BEEBE, W.
don, vol. 139, pt. 3, pp. 451-473. 1934. Deep-sea fishes of the Bermuda ocean-
1963. Frontal foramina and tripodes of the ographic expeditions. Family Idiacan-
characin Crenuchus. Nature, vol. 200, thidae. Zoologica, vol. 16, no. 4, pp. 148-
no. 4912, p. 1225. 241.
1964a. The structure of the Weberian appara- BEEBE, W., AND J. CRANE
tus in the Siluri. Proc. Zool. Soc. Lon- 1939. Deep-sea fishes of the Bermuda ocean-
don, vol. 142, pt. 3, pp. 419-440. ographic expeditions. Family Melano-
1964b. The evolution of the Weberian appara- stomiatidae. Zoologica, vol. 24, pt. 2,
tus in the Cobitidae. Ibid., vol. 143, pt. no. 6, pp. 65-238.
1, pp. 177-190. BERG, L. S.
ALLIS, E. P. 1940. Classification of fishes, both Recent and
1898. On the morphology of certain of the fossil. Trav. Inst. Zool. Acad. Sci.
bones of the cheek and snout of Amia U.S.S.R., vol. 5, no. 2, 517 pp. (Russian
calva. Jour. Morph., vol. 14, pp. 425- and English texts. Also reprint, Ann
466. Arbor, Michigan, 1947.)
1903. The lateral sensory system in the Mur- 1955. Systema ribobraznich i rib nine jivucht-
aenidae. Internatl. Monatsschr. Anat. chich i iskopaemich. Trudy Zool. Inst.
und Physiol., vol. 20, pp. 125-170. Akad. Nauk S.S.S.R., vol. 20, pp. 1-
1909. The cranial anatomy of the mail- 286.
cheeked fishes. Zoologica, Stuttgart, vol. 1958. System der rezenten und fossilen Fisch-
22, no. 57, 219 pp. artigen und Fische. Berlin, xii+311 pp.
ARAMBOURG, C. BERTELSEN, E., AND N. B. MARSHALL
1954. Les poissons cretaces du Jebel Tselfat 1956. The Miripinnati, a new order of teleost
(Maroc). Notes et MWm. Serv. Gaol. fishes. In Dana-Report. Copenhagen,
Maroc, Rabat, no. 118, pp. 1-188. no. 42, 34 pp.
D'AusENTON, F. BERTIN, L., AND C. ARAMBOURG
1955. ttude de l'appareil branchiospinal et de 1958. Super-ordre des t6leosteens (Teleostei).
l'organe suprabranchial d'Heterotis nilo- In Grasse, P., Traite de zoologie. Paris,
ticus Ehrenberg 1827. Bull. l'Inst. Fran- vol. 13, fasc. 3, pp. 2204-2500.
gais d'Afrique Noire, ser. A, vol. 17, no. BERTMAR, G.
4, pp. 1179-1201. 1959. On the ontogeny of the chondral skull in
1961. Morphologie du crAne de Cromeria Characidae, with a discussion on the
nilotica occidentalis Daget 1954. Ibid., chondrocranial base and the visceral
ser. A, vol. 23, no. 1, pp. 134-163. chondrocranium in fishes. Act. Zool.
BAILEY, R. M. Stockholm, vol. 40, pp. 203-364.
1960. Forty-five articles on Recent fishes. 1962. On the ontogeny and evolution of the
Reprinted from the McGraw-Hill En- arterial vascular system in the head of
cyclopedia of Science and Technology. the African characidean fish Hepsetus
New York, McGraw-Hill Book Co., odoe. Ibid., vol. 43, pp. 255-295.
Inc., 21 pp. B6HLKE, J.
BAMFORD, T. W. 1956. A synopsis of the eels of the family Xen-
1948. The cranial development of Galeichthys ocongridae (including the Chlopsidae
felis. Proc. Zool. Soc. London, vol. 118, and Chilorhinidae). Proc. Acad. Nat.
no. 2, pp. 364-391. Sci. Philadelphia, vol. 108, pp. 61-95.
BARDACK, D. 1960. Comments on serranoid fishes with dis-
1964. A revision of fossil and living chirocent- junct lateral lines, with the description
rid fishes. Ph.D., University of Kansas, of a new one from the Bahamas. Notulae
1963. Diss. Abstr., Ann Arbor, Mich- Nat., no. 330, pp. 1-11.
igan, vol. 24, no. 12, pt. 1, pp. 5618- BOULENGER, G. A.
5619. 1898. A revision of the genera and species of
BEAUFORT, L. F. DE fishes of the family Mormyridae. Proc.
1909. Die Schwimmblase der Malacopterygii. Zool. Soc. London, pp. 775-821.
436
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 437
1901. Notes on the classification of teleostean myrus rume C. et V. Ibid., ser. A, vol.
fishes. I. On the Trachinidae and their 22, no. 3, pp. 1013-1052.
allies. Ann. Mag. Nat. HIist., ser. 7, vol. DARLINGTON, P. J., JR.
8, no. 46, pp. 261-271. 1957. Zoogeography: the geographical dis-
1904. Fishes. (Systematic account of Tele- tribution of animals. New York, John
ostei.) In Harmer, S. F., and A. E. Wiley and Sons, Inc., xi+675 pp.
Shipley (eds.), The Cambridge natural DAVID, L.
history. London, Macmillan and Co., 1935. Die Entwicklung der Clariiden und
Ltd., vol. 7, pp. 539-727. ihre Verbreitung. Eine anatomisch-
BRIDGE, T. W., AND A. C. HADDON systematische Untersuchung. Rev. Zool.
1893. Contribution to the anatomy of fishes. Bot. Africaines, vol. 28, no. 1, pp. 77-
II. The air-bladder and Weberian ossi- 147.
cles in the siluroid fishes. Phil. Trans. DEHADRAI, P. V.
Roy. Soc. London, ser. B, vol. 84, pp. 1957. On the swimbladder and itsrelation with
65-333. the internal ear in genus Notopterus
BRITTAN, M. R. (Lac6pede). Jour. Zool. Soc. India, vol.
1954. A revision of the Indo-Malayan fresh- 9, pp. 50-61.
water fish genus Rasbora. Monogr. Inst. EATON, T. H.
Sci. and Tech., Manila, vol. 3, 224 pp. 1948. Form and function in the head of the
CHRANILOV, N. S. channel catfish (Ictalurus lacustris punc-
1929. Beitrage zur Kenntnis des Weber'schen tatus). Jour. Morph., vol. 83, pp. 181-
Apparates der Ostariophysi. 2. Der 194.
Weber'schen Apparat bei Siluroidea. EGE, V.
Zool. Jahrb., Anat., vol. 51, pp. 323- 1934. The genus Stomias Cuv., taxonomy and
462. bio-geography. In Dana-Report. Copen-
CLAUSEN, H. S. hagen, no. 5, 58 pp.
1959. Denticipitidae, a new family of primi- 1948. Chauliodus Schn., bathypelagic genus
tive isospondylous teleosts from west of fishes. A systematic, phylogenetic and
African fresh-water. Vidensk. Meddel. geographic study. Op. cit. Copenhagen,
fra Dansk naturhist. Foren., vol. 121, no. 31, 148 pp.
pp. 141-156. EIGENMANN, C. H.
COCKERELL, T. D. A. 1912. The fresh-water fishes of British Guiana.
1910a. On the scales of some malacopterygian Mem. Carnegie Mus., vol. 5, xxii
fishes. Proc. Biol. Soc. Washington, vol. +578 pp.
23, pp. 111-114. 1927. The fresh-water fishes of Chile. Mem.
1910b. The scales of the mormyroid fishes with Natl. Acad. Sci., vol. 22, no. 2, pp. 1-63.
remarks on Albula and Elops. Smith- EIGENMANN, C. H., AND W. R. ALLEN
sonian Misc. Coll., vol. 56, no. 3, pp. 1-4. 1942. Fishes of western South America. Lex-
COPE, E. D. ington, University of Kentucky, 494 pp.
1871. Contribution to the ichthyology of the ELLIS, M. M.
Lesser Antilles. Trans. Amer. Phil. Soc., 1913. Gymnotid eels of tropical America.
new ser., vol. 14, pp. 445-483. Mem. Carnegie Mus., vol. 6, no. 3,
COUCEIRO, A., AND D. F. DE ALMEIDA pp. 109-195.
1961. The electrogenic tissue of some Gym- EMERY, C.
notidae. In Chagas, C., and A. Paes de 1880. Fierasfer. Studi intorno alla sistematica,
Carvalho (eds.), Bioelectrogenesis. A l'anatomia e la biologia delle specie
comparative survey of its mechanisms mediterranee di questo genere. Atti R.
with particular emphasis on electric Accad. Lincei, ser. 3, vol. 7, pp. 167-254.
fishes. Proc. Symposium on Comp. Bio- FORSKXL, P.
electrogenesis, Rio de Janeiro, 1959, 1775. Descriptiones animalium, avium, am-
Amsterdam, London, New York, pp. 3- phibiorum, piscium, insectorum, ver-
13. mium. Copenhagen, 20+xxxiv+ 164 pp.
DAGET, J., AND F. D'AUBENTON FRANZ, V.
1957. D6veloppement et morphologie du crine 1920. Einige Hauptpunkte in der Organisation
d'Heterotis niloticus Ehr. Bull. l'Inst. der Mormyriden. Biol. Zentralbl., vol.
Frangais d'Afrique Noire, ser. A, vol. 40, no. 1, pp. 1-15.
19, no. 3, pp. 881-936. FRIEHIORER, W. C.
1960. Morphologie du chondrocrAne de Mor- 1963. Patterns of the ramus lateralis acces-
438 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
sorius and their systematic significance 1961. Some osteological features of modern
in teleostean fishes. Stanford Ichthyol. lower teleostean fishes. Smithsonian
Bull., vol. 8, no. 2, pp. 80-189. Misc. Coll., vol. 142, no. 3, pp. 1-42.
GARSTANG, W. GREENWOOD, P. H.
1931. The phyletic classification of Teleostei. 1960. Fossil denticipitid fishes from east
Proc. Leeds Phil. Lit. Soc., sci. sect., Africa. Bull. Brit. Mus. (Nat. Hist.),
vol. 2, pt. 5, pp. 240-260. geol., vol. 5, no. 1, pp. 1-11.
GERY, J. 1963. The swimbladder in African Noto-
1964. Une nouvelle famille de poissons dul- pteridae (Pisces) and its bearing on the
caquicoles africains: les Grasseichthy- taxonomy of the family. Ibid., zool.,
idae. Compt. Rendus Acad. Sci., Paris, vol. 11, no. 5, pp. 377-412.
vol. 259, no. 25, pp. 4805-4807. 1965. The status of Acanthothrissa Gras, 1961
GERY, J., AND VU-TAN-TUE (Pisces, Clupeidae). Ann. Mag. Nat.
1964. Gymnorhamphichthys hypostomus pe- Hist., ser. B, vol. 7, pp. 337-338.
titi, ssp. nov., un curieux poisson gym- GREENWOOD, P. H., AND K. S. THOMSON
notoide arenecole. Vie et Milieu, Vol. 1960. The pectoral anatomy of Pantodon
Jubilaire dedi6 a Georges Petit, suppl. buchholzi Peters (a freshwater flying fish)
no. 17, pp. 485-498. and the related Osteoglossidae. Proc.
GILL, T. N. Zool. Soc. London, vol. 135, pp. 283-301.
1872. Arrangement of the families of fishes, or GREGORY, W. K.
classes Pisces, Marsipobranchii, and 1907. The orders of teleostomous fishes. A pre-
Leptocardii. Smithsonian Misc. Coll., liminary review of the broader features
vol. 11, no. 247, xlvi+49 pp. of their evolution and taxonomy. Ann.
1884. On the anacanthine fishes. Proc. Acad. New York Acad. Sci., vol. 17, pt. 2,
Nat. Sci. Philadelphia (1883), pp. 167- no. 3, pp. 437-508.
183. 1933. Fish skulls: a study of the evolution of
1893. Families and subfamilies of fishes. natural mechanisms. Trans. Amer. Phil.
Mem. Natl. Acad. Sci., vol. 6, pp. 127- Soc., new ser., vol. 23, art. 2, pp. 75-481.
138. GREGORY, W. K., AND G. M. CONRAD
GILTAY, L. 1936. Pictorial phylogenies of deep sea Iso-
1934. Note ichthyologique IX. Contribution spondyli and Iniomi. Copeia, no. 1,
A l'6tude du genre Xenopomatichthys pp. 21-36.
(Kneriidae). Bull. Mus. Roy. Hist. Nat. GREY, M.
Belgique, vol. 10, no. 44, pp. 1-22. 1960. A preliminary review of the family
GOLVAN, Y.-J. Gonostomatidae, with a key to the
1962. Catalogue systematique des noms de genera and the description of a new
genres de poissons actuels de la XI species from the tropical Pacific. Bull.
Edition du "Systema naturae" de Charles Mus. Comp. Zool. Harvard, vol. 122,
Linn6 jusqu'a la fin de l'annee 1959. no. 2, pp. 57-125.
Ann. Parasitol. Humaine et Comp., GtNTHER, A.
vol. 37, no. 6 bis, fasc. suppl., pp. 1-227. 1862. Acanthopterygii Pharyngognathi and
GOODRICH, E. S. Anacanthini. In Gunther, A., Catalogue
1909. Cyclostomes and fishes. In Lankester, of the fishes in the British Museum.
R. (ed.), A treatise on zoology. Pt. 9. London, vol. 4, 534 pp.
London, Adam and Charles Black, 1868. Physostomi. In Gunther, A., op. cit.
xvi+518 pp. London, vol. 7, 512 pp.
GOSLINE, W. A. GtNTHER, K., AND K. DECKERT
1952. Notes on the systematic status of four 1955. Zweiter Versuch einer morphologisch-
eel families. Jour. Washington Acad. anatomischen Funktionanalyse der
Sci., vol. 42, no. 4, pp. 130-135. Nahrungserwerbs- und Atmungsappa-
1959. Mode of life, functional morphology, rate von Tiefseefischen. Zool. Beitr.,
and the classification of modern teleos- new ser., vol. 1, no. 3, pp. 241-365.
tean fishes. Syst. Zool., vol. 8, no. 3, 1959. Morphologie und Funktion des Kiefer-
pp. 160-164. und Kiemenapparates von Tiefseefisch-
1960. Contributions toward a classification of en der Gattungen Malacosteus und
modern isospondylous fishes. Bull. Brit. Photostomias (Teleostei, Isospondyli,
Mus. (Nat. Hist.), zool., vol. 6, no. 6, Stomiatoidea, Malacosteidae). In Dana-
pp. 325-365. Report. Copenhagen, no. 49, 54 pp.
966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 439
IARRINGTON, R. W. Proc. Natl. Inst. Sci. India, ser. B,
1955. The osteocranium of the American vol. 26, pp. 205-233.
cyprinid fish, Notropis bifrenatus, with KINDRED, J. E.
an annotated synonymy of teleost skull 1919. The skull of Ameiurus. Illinois Biol.
bones. Copeia, no. 4, pp. 267-296. Monogr., vol. 5, pp. 1-120.
IAY, 0. P. KNER, R.
1903. On a collection of Upper Cretaceous 1869. Fische. In Reise der Osterreichischen
fishes from Mount Lebanon, Syria, with Fregatte Novara um die Erde in den
descriptions of four new genera and Jahren 1857, 1858, 1859, B. von Wijiler-
nineteen new species. Bull. Amer. Mus. storf-Urbair. Vienna, Zoologischer Theil,
Nat. Hist., vol. 19, art. 10, pp. 395-452. vol. 1, no. 5, pp. 1-433.
IILLS, E. S. KRANDIKAR, K. R., AND V. B. MASUREKAR
1934. Tertiary fresh water fishes from south- 1954. Weberian apparatus and other related
ern Queensland. Mem. Queensland Mus., structures of Arius platystomus Day.
Brisbane, vol. 10, pp. 157-174. Jour. Bombay Univ., vol. 22, no. 5,
IOEDEMAN, J. J. pp. 1-28.
1954. Aquariumvissen-encyclopaedie. Amster- LE DANOIS, E.
dam, 577 pp. 1943. Remarques ichthyologiques. Rev. Trav.
1960. Remarks on the classification and phy- l'Office Sci. Tech. P6ches Maritimes,
logeny of teleostean fishes. Bull. Aquatic vol. 13, fascs. 1-4, nos. 49-52, pp. 54-
Biol., vol. 2, no. 15, pp. 50-51. 175.
LE DANOIS, Y.
IOLLISTER, G. 1961. Remarques sur les poissons orbiculates
1936. Caudal skeleton of Bermuda shallow du sous-ordre des Ostracioniformes.
water fishes. I. Order Isospondyli: Mem. Mus. Natl. Hist. Nat., Paris, new
Elopidae, Megalopidae, Albulidae, Clu- ser., ser. A, zool., vol. 19, fasc. 2, pp.
peidae, Dussumieriidae, Engraulidae. 207-338.
Zoologica, vol. 21, pt. 4, no. 23, pp. 257- LERAY, C.
290. 1961. Contribution a l'6tude osteologique de
IUBBS, C. L. Gouania wildenowii Risso (tWl6ost6ens)
1945. Phylogenetic position of the Citharidae, (squelette cephalique et ceintures).
a family of flatfishes. Misc. Publ. Mus. Cahiers Biol. Marine, vol. 2, no. 1,
Zool. Univ. Michigan, no. 63, pp. 1-38. pp. 41-52.
1953. Synonymy of the bathypelagic fish LIEM, K. F.
genus Rhyncohoyalus, referred to the ex- 1963. The comparative osteology and phy-
panded family Argentinidae. Copeia, logeny of the Anabantoidei (Teleostei,
no. 2, pp. 96-97. Pisces). Illinois Biol. Monogr., no. 30,
AYARAM, K. C. 149 pp.
1955. Taxonomic status of Chinese catfish of LISSMANN, H. W.
the family Cranoglanidae. Proc. Natl. 1961. Ecological studies on gymnotids. In
Inst. Sci. India, ser. B, vol. 21, pp. 256- Chagas, C., and A. Paes de Carvalho
263. (eds.), Bioelectrogenesis. A comparative
ORDAN, D. S. survey of its mechanisms with particular
1905. A guide to the study of fishes. New York, emphasis on electric fishes. Proc. Sym-
Henry Holt and Co., vol. 2, xxii+599 posium on Comp. Bioelectrogenesis,
Rio de Janeiro, 1959, Amsterdam,
pp- London, New York, pp. 215-226.
1923. Classification of fishes including families MCMURRICH, J. P.
and genera as far as known. Stanford 1884. The osteology of Ameiurus catus (L.)
Univ. Publ., Univ. Ser., Biol. Sci., vol. Gill. Proc. Canadian Inst., Toronto,
3, pp. 77-243. new ser., vol. 2, pp. 270-310.
rORDAN, D. S., AND B. W. EVERMANN MAKUSHOK, V. M.
1898. The fishes of North and Middle Amer- 1961. The place of Neozoarcinae (Zoarcidae,
ica. Bull. U. S. Natl. Mus., no. 47, pt. 3, Blennioidei, Pisces), in classification of
pp. 2183-3136. fishes. Akad. Nauk, U.S.S.R., vol. 43,
fOSEPH, N. I. pp. 198-224. (In Russian.)
1960. Osteology of Wallago attu Bloch and MARCK, W. VON DER
Schneider. Pt. 1. Osteology of head. 1858. Vber einige Wirbelthiere, Kruster und
AAn BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
Cephalopoden der Westfalischen Kreide. Ailia coila (Ham.) during growth. Rec.
Zeitschr. Deutschen Geol. Gesell., Ber- Indian Mus., vol. 40, pp. 183-187.
lin, vol. 10, pp. 231-271. NAWAR, G.
1863. Fossile Fische, Krebse und Pflanzen aus 1954. On the anatomy of Clarias lazera.
dem Plattenkalk der jungsten Kreide in I. Osteology. Jour. Morph., vol. 94,
Westphalen. Palaeontographica, vol. 11, pp. 551-585.
nos. 1-2, pp. 1-83. NELSON, E. M.
MARSEALL, N. B. 1948. The comparative morphology of the
1962. Observations on the Heteromi, an order Weberian apparatus of the Catostomi-
of teleost fishes. Bull. Brit. Mus. (Nat. dae and its significance in systematics.
Hist.), zool., vol. 9, no. 6, pp. 249-270. Jour. Morph., vol. 83, pp. 225-251.
MASUREKAR, V. B. NICHOLS, J. T.
1962. Weberian apparatus in Mystus gulio 1930. Speculation on the history of the Ostari-
(Ham.-Buch.) and Mystus bleekeri ophysi. Copeia, no. 4, pp. 148-151.
(Day). Proc. Indian Acad. Sci., ser. B, 1943. The fresh-water fishes of China. Natural
vol. 55,
no. 1, pp. 24-37. history of central Asia, vol. 9. New
MATSUBARA, K. York, the American Museum of Natural
1943. Studies on the scorpaenoid fishes. Pts. History, xxxvi+322 pp.
I and II. Trans. Sigenkagaku Ken- NORMAN, J. R.
kyusyo (Res. Inst. Nat. Resources, 1930. Oceanic fishes and flatfishes collected in
Tokyo), nos. 1 and 2, 486 pp. 1925-1927. Discovery Repts., no. 2,
1955. Fish morphology and hierarchy. Pts. I- pp. 261-370.
III. Tokyo, Ishizaki-Shoten, 1605 pp., NYBELIN, 0.
135 pls. 1957. Les canaux sensorials du museau chez
1963. Pisces. In Uchida, T. (ed), Systematic Elops saurus (L.). Notice preliminaire.
zoology. Tokyo, Nakayama Shoten, Arkiv f. Zool., vol. 10, pp. 453-458.
vol. 9, p. 2, pp. 197-531. 1961. ttber die Frage der Abstammung der
MEAD, G. W. rezenten primitiven Teleostier. Palaont.
1965. The larval form of the Heteromi Zeitschr., vol. 35, pp. 114-117.
(Pisces). Breviora, no. 226, pp. 1-5. 1964. Versuch einer taxonomischen Revision
MERRIMAN, D. der jurassischen Fischgattung Thrissops
1940. Morphological and embryological stud- Agassiz. Goteborgs K. Vetensk. Vitter-
ies on two species of marine catfish, hetssamhalles Handl., ser. B, vol. 9, pt.
Bagre marinus and Galeichthys felis. 4, pp. 1-44.
Zoologica, vol. 25, no. 13, pp. 221-248. NYSTEN, M.
M#LLER, J. 1962. etude anatomique des rapports de la
1845. tVber den Bau und die Grenzen der vessie a&rienne avec 1'axe vertebral chez
Ganoiden und uiber das natiirlichen Pantodon buchholzi Peters. Ann. Mus.
System der Fische. Abhandl. Akad. Roy. l'Afrique Centrale, Tervuren,
Wiss. Berlin, 1844, pp. 117-216. Belgique, ser. 8 (zool. sci.), no. 108,
MYERS, G. S. pp. 187-220.
1938. Fresh-water fishes and West Indian OMARKHAN, M.
zoogeography. Ann. Rept. Smithsonian 1949a. The morphology of the chondrocranium
Inst., for 1937, pp. 339-364. of Gymnarchus niloticus. Jour. Linnean
1958. Trends in the evolution of teleostean Soc. London, Zool., vol. 61, pp. 452-481.
fishes. Stanford Ichthyol. Bull., vol. 7, 1949b. The lateral sensory canals of larval
no. 3, pp. 27-30. Notopterus. Proc. Zool. Soc. London,
[In press.] Fresh-water fishes and continental vol. 118, pt. 4, pp. 938-970.
drift. Proc. California Acad. Sci. 1950. The development of the chondro-
NABLANT, T. cranium of Notopterus. Jour. Linnean
1963. A study of the genera of Botiinae and Soc. London, Zool., vol. 61, pp. 608-624.
Cobitinae (Pisces, Ostariophysi, Co- ORTON, G. L.
bitidae). Tray. Mus. Hist. Nat. "Grigore 1963. Notes on larval anatomy of fishes of the
Antipa," Bucharest, vol. 4, pp. 343-379. order Lyomeri. Copeia, no. 1, pp. 6-15.
NAIR, K. K. PARR, A. E.
1938. Changes in the internal structure of 1927. The stomiatoid fishes of the suborder
the airbladder of Eutropiichthys vacha Gymnophotodermi (Astronesthidae,
(Ham.), Clupisoma garua (Ham.) and Melanostomiatidae, Idiacanthidae) with
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 441
a complete review of the species. Bull. to phylogenetic studies. 5. The skull and
Bingham Oceanogr. Coll., vol. 3, art. gasbladder capsule of the Cobitidae.
2, 123 pp. Ibid., vol. 19, no. 3, pp. 323-347.
1930. A note on the classification of the 1955a. Skeleton of cyprinoid fishes in relation
stomiatoid fishes. Copeia, no. 4, p. 136. to phylogenetic studies. 6. The skull and
1937. Concluding report on fishes. Bull. Bing- Weberian apparatus in the subfamily
ham Oceanogr. Coll., vol. 3, pp. 1-79. Gobioninae (Cyprinidae). Acta Zool.,
1951. Preliminary revision of the Alepocephal- Stockholm, vol. 36, pp. 127-158.
idae, with the introduction of a new 1955b. Skeleton of cyprinoid fishes in relation
family, Searsidae. Amer. Mus. Novi- to phylogenetic studies. 7. The skull and
tates, no. 1531, pp. 1-21. Weberian apparatus of Cyprininae (Cy-
1952. A revision of the species currently re- prinidae). Ibid., vol. 36, pp. 199-242.
ferred to Alepocephalus, Halisauriceps, 1957. Skeleton of cyprinoid fishes in relation
Bathytroctes, and Bajacalifornia with to phylogenetic studies. 8. The skull and
introduction of two new genera. Bull. Weberian ossicles of Catostomidae.
Mus. Comp. Zool., Harvard, vol. 107, Proc. Zool. Soc., Calcutta, Mookerjee
pp. 253-269. Mem. Vol., pp. 293-303.
1960. The fishes of the family Searsidae. In RAYNER, D. H.
Dana-Report. Copenhagen, no. 51, 1936. On Leptolepis bronni Agassiz. Ann. Mag.
pp. 1-108. Nat. Hist., ser. 10, vol. 11, pp. 46-74.
PATTERSON, C. RECHNITZER, A. B., AND J. B6HLKE
1964. A review of Mesozoic acanthopterygian 1958. Ichthyococcus irregularis, a new gono-
fishes, with special reference to those of stomatine fish from the eastern Pacific.
the English Chalk. Phil. Trans. Roy. Copeia, no. 1, pp. 10-15.
Soc. London, ser. B, biol. sci., vol. 247, REGAN, C. T.
no. 739, pp. 213-482. 1903. On th,e systematic position and classi-
PITON, L. fication of the gadoid or anacanthine
1938. Les Characinidae fossiles de Menat fishes. Ann. Mag. Nat. Hist., ser. 7,
(P.-de-D.). Rev. Sci. Bourbonnais Cent. vol. 11, pp. 459-466.
France, Moulins, no. 3-4, pp. 98-104. 1909. The classification of teleostean fishes.
QUAST, J. C. Ibid., ser. 8, vol. 3, pp. 75-86.
1965. Osteological characteristics and affinities 1911a. The classification of the teleostean fishes
of the hexagrammid fishes, with a synop- of the order Ostariophysi. 1. Cyprinoi-
sis. Proc. California Acad. Sci., ser. 4, dea. Ibid., ser. 8, vol. 8, pp. 13-32.
vol. 31, no. 21, pp. 563-600. 1911b. The classification of the teleostean fishes
RAMASWAMI, L. S. of the order Ostariophysi. 2. Siluroidea.
1948. The homalopterid skull. Proc. Zool. Soc. Ibid., ser. 8, vol. 8, pp. 553-577.
London, vol. 118, pt. 2, pp. 515-538. 1912a. The classification of the teleostean fishes
1952a. Skeleton of cyprinoid fishes in relation of the order Pediculati. Ibid., ser. 8,
to phylogenetic studies. 1. The syste- vol. 9, pp. 277-289.
matic position of the genus Gyrino- 1912b. The classification of the blennioid fishes.
cheilus Vaillant. Proc. Natl. Inst. Sci. Ibid., ser. 8, vol. 10, pp. 265-280.
India, vol. 18, no. 2, pp. 125-140. 1912c. The anatomy and classification of the
1952b. Skeleton of cyprinoid fishes in relation teleostean fishes of the order Lyomeri.
to phylogenetic studies. 2. The syste- Ibid., ser. 8, vol. 10, pp. 347-349.
matic position of Psilorhynchus Mc- 1922. The distribution of the fishes of the
Clelland. Ibid., vol. 18, no. 2, pp. 141- order Ostariophysi. Bijd. tot de Dierk.,
150. Amsterdam, vol. 22, pp. 203-207.
1952c. Skeleton of cyprinoid fishes in relation 1923. The classification of the stomiatoid
to phylogenetic studies. 3. The skull and fishes. Ann. Mag. Nat. Hist., ser. 9,
other skeletal structures of homalopterid vol. 11, pp. 612-614.
fishes. Ibid., vol. 18, no. 6, pp. 495-517. 1929. Fishes. In Encyclopaedia Britannica.
1952d. The skeleton of cyprinoid fishes in re- Fourteenth edition. London and New
lation to phylogenetic studies. 4. The York, vol. 9, pp. 305-329.
skull and other skeletal structures of REGAN, C. T., AND E. TREWAVAS
gastromyzonid fishes. Ibid., vol. 18, 1929. The fishes of the families Astrones-
no. 6, pp. 519-538. thidae and Chauliodontidae. Dana Ex-
1953. Skeleton of cyprinoid fishes in relation ped. 1920-22 Rept., no. 5, 39 pp.
442 BULLETIN AMERICAN MU'SE;UM OF NATURAL HISTORY VOL. 131
1930. The fishes of the families Stomiatidae SCHULTZ, L. P.
and Malacosteidae. Dana Exped. 1920- 1961. Revision of the marine silver hatchet-
22 Rept., no. 6, 143 pp. fishes (family Sternoptychidae). Proc.
REICHEL, M. U. S. Natl. Mus., vol. 112, no. 3449,
1927. etude anatomique du Phreatobius cis- pp. 587-649.
ternarum Goeldi, silure aveugle du SHELDEN, E. E.
Br6sil. Rev. Suisse Zool., vol. 34, no. 16, 1937. Osteology, myology and probable evo-
pp. 285-403. lution of the nematognath pelvic girdle.
RENSCH, B. Ann. New York Acad. Sci., vol. 37, no. 1,
1960. Evolution above the species level. New pp. 1-96.
York, Columbia University Press, xviii SIMPSON, G. G.
+419 pp. 1945. The principles of classification and a
RICRARDSON, J. classification of mammals. Bull. Amer.
1843. Contributions to the ichthyology of Mus. Nat. Hist., vol. 85, xvi+350 pp.
Australia. Ann. Mag. Nat. Hist., vol. 11, 1961. Principles of animal taxonomy. New
no. 73-suppl., art. 63, pp. 489-498. York, Columbia University Press, xii
RIDEWOOD, W. G. +247 pp.
1904a. On the cranial osteology of the fishes of SMITH, C. L.
the families Elopidae and Albulidae, 1965. The patterns of sexuality and the classi-
with remarks on the morphology of the fication of serranid fishes. Amer. Mus.
skull in the lower teleostean fishes Novitates, no. 2207, pp. 1-20.
generally. Proc. Zool. Soc. London, vol. SMITH, C. L., AND R. M. BAILEY
2, pp. 35-81. 1962. The subocular shelf of fishes. Jour.
1904b. On the cranial osteology of the fishes of Morph., vol. 110, no. 1, pp. 1-18.
the families Mormyridae, Notopteridae, SRINIVASACHAR, H. R.
and Hyodontidae. Jour. Linnean Soc. 1957. Development of the skull in catfishes.
London, Zool., vol. 29, pp. 188-217. II. Development of chondrocranium in
1905a. On the cranial osteology of the fishes of Mystus and Rita (Bagridae). Morph.
the families Osteoglossidae, Pantodonti- Jahrb., vol. 98, no. 2, pp. 224-262.
dae and Phractolaemidae. Ibid., vol. 29- 1958. Development of skull in catfishes. 5. De-
pp. 252-282. velopment of skull in Heteropneustes
1905b. On the cranial osteology of the clupeoid fossilis Bloch. Proc. Natl. Inst. Sci.
fishes. Proc. Zool. Soc. London, vol. 2, India, ser. B, vol. 24, pp. 165-190.
pp. 448-493. 1959. Development of the skull in catfishes.
1905c. On the skull of Gonorkynchus Greyi. IV. The development of chondro-
Ann. Mag. Nat. Hist., ser. 7, vol. 15, cranium in Arius jella Day (Ariidae)
no. 88, art. 45, pp. 361-372. and Plotosus canius (Plotosidae) with
ROSEN, D. E. account of their inter-relationship.
1962. Comments on the relationships of the Morph. Jahrb., vol. 99, no. 4, pp. 986-
North American cave fishes of the 1016.
family Amblyopsidae. Amer. Mus. No- STARKS, E. C.
vitates, no. 2109, pp. 1-35. 1905. The osteology of Caularchus maeandri-
1964. The relationships and taxonomic posi- cus (Girard). Biol. Bull., vol. 9, no. 5,
tion of the halfbeaks, killifishes, silver- pp. 292-303.
1923. The osteology and relationships of the
sides, and their relatives. Bull. Amer. uranoscopoid fishes, with notes on other
Mus. Nat. Hist., vol. 127, art. 5, fishes with jugular ventrals. Stanford
pp. 217-268. Univ. Publ., Univ. Ser., Biol. Sci., vol.
SAINT-SEINE, P. DE 3, no. 3, pp. 259-290.
1949. Les poissons des calcaires lithograph- 1930. The primary shoulder girdle of the bony
iques de Cerin (Ain). Nouv. Arch. Mus. fishes. Ibid., Univ. Ser., Biol. Sci., vol.
d'Hist. Nat. Lyon, fasc. 2, pp. 1-357. 6, pp. 147-239.
SANDERS, M. STIPETI6, E.
1934. Die fossilen Fische der alttertiiaren 1939. tJber das Geh6rorgan der Mormyriden.
Siisswasserablagerungen aus Mittel- Zeitschr. f. Vergl. Physiol., vol. 26,
Sumatra. Verhandel. Geol.-Mijn. Ge- pt. 5, pp. 740-752.
nootsch. Nederland en Kolonien, geol. SVETOVIDOV, A. N.
ser., vol. 11, pp. 1-143. 1948. Gadiformes. In Pavlovskii, E. N., and
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 443
A. A. Shtakel'berg (eds.), Fauna of the Zool. Anz., vol. 171, nos. 11-12, pp. 424-
U.S.S.R., Fishes, vol. 9, no. 4. Zool. 439.
Inst. Akad. Nauk SSSR, new ser., 1964. The osteocranium and the Weberian
no. 34. (Translation from the Russian; apparatus of the fishes of the family
published for the National Science Schilbeidae (Pisces: Siluroidea). Proc.
Foundation, Washington, D. C., and the Zool. Soc. London, vol. 143, pt. 1,
Smithsonian Institution, 1962, 304 pp.) pp. 1-36.
1961. On European species of fam. Ophidiidae 1965. The comparative morphology of the
and on the functional importance of the osteocranium and the Weberian appar-
structure specificity of their fish-sound. atus of Tachysuridae (Pisces: Silu-
Voprosy Ikhtiologii, vol. 17, pp. 3-13. roidei). Jour. Zool., vol. 146, pt. 2, pp.
(In Russian.) 150-174.
SWINNERTON, H. H. TRETIAKOV, D. K.
1903. The osteology of Cromeria nilotica and 1944. Outlines of the phylogeny of fishes.
Galaxias attenuatus. Zool. Jahrb., Jena, [Leningrad?], Akademiia Nauk SSSR,
vol. 18, no. 1, pp. 58-70. Zoologicheskii Institute, 176 pp. (In
TCEERNAVIN, V. V. Russian, translated from the Ukrain-
1947a. Six specimens of Lyomeri in the British ian.)
Museum (with notes on the skeleton of TREWAVAS, E.
Lyomeri). Jour. Linnean Soc. London, 1933. On the structure of two oceanic fishes,
Zool., vol. 41, pp. 287-350. Cyema atrum Gurnther, and Opistho-
1947b. Further notes on the structure of the proctus soleatus Vaillant. Proc. Zool. Soc.
bony fishes of the order Lyomeri London, pp. 601-614.
(Eurypharynx). Ibid., Zool., vol. 41, VALENCIENNES, M. A.
pp. 377-393. 1846. Histoire naturelle des poissons. Paris,
1953. The feeding mechanisms of a deep sea P. Bertrand, vol. 19, 534 pp.
fish, Chauliodus sloani Schneider. Lon- WEITZMAN, S. H.
don, British Museum (Natural History), 1954. The osteology and the relationships of
101 pp. the South American characid fishes of
THIELE, H. the subfamily Gasteropelecinae. Stan-
1963. Vergleichend-morphologische Unter- ford Ichthyol. Bull., vol. 4, no. 4,
suchungen iiber die Funktion der Nahr- pp. 213-263.
ungserwerbsapparate von Anarrkichas 1960a. The systematic position of Piton's pre-
lupus L. und einigen Blenniidae (Tele- sumed characid fishes from the Eocene
ostei, Blennioidei). Zool. Beitr., new ser., of central France. Ibid., vol. 7, no. 4,
vol. 9, nos. 2 and 3, pp. 275-440. pp. 114-123.
THYS VAN DEN AUDENAERDE, D. F. E. 1960b. Further notes on characid fossils. Ibid.,
1961. L'anatomie de Phractolaemus ansorgei vol. 7, no. 4, pp. 215-216.
Blgr. et la position syst6matique des 1960c. Further notes on the relationships and
Phractolaemidae. Ann. Mus. Roy. l'Af- classification of the South American
rique Centrale, Tervuren, ser. 8, zool. characid fishes of the subfamily Gas-
sci., no. 103, pp. 101-167. teropelecinae. Ibid., vol. 7, no. 4,
TILAK, R. pp. 217-239.
1963a. Studies on the nematognathine pectoral 1962. The osteology of Brycon meeki, a gener-
girdle in relation to taxonomy. Ann. alized characid fish, with an osteological
Mag. Nat. Hist., ser. 13, vol. 6, pp. 145- definition of the family. Ibid., vol. 8,
155. no. 1, pp. 1-77.
1963b. The osteocranium and the Weberian ap- 1964. Osteology and relationships of South
paratus of the fishes of the family American characid fishes of subfamilies
Sisoridae (Siluriodea): a study in adap- Lebiasininae and Erythrininae, with
tation and taxonomy. Zeitschr. Wiss. special reference to subtribe Nanno-
Zool. Leipzig, vol. 168, nos. 3-4, stomina. Proc. U. S. Natl. Mus., vol.
pp. 281-320. 116, no. 3499, pp. 127-169.
1963c. The osteocranium and Weberian appa- WHITEHEAD, P. J. P.
ratus of a few representatives of the 1963. A contribution to the classification of
families Siluridae and Plotosidae (Silu- clupeoid fishes. Ann. Mag. Nat. Hist.,
roidea): a study of inter-relationship. ser. 13, vol. 5, pp. 737-750.
444 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
WOHLFAHRT, T. A. British Museum. London, pt. 4, 636 pp.
1936. Das Ohrlabyrinth der Sardine (Clupea 1942. The beginning of the teleostean fishes.
pilchardus Walb.) und seine Beziehun- Ann. Mag. Nat. Hist., ser. 11, vol. 9,
gen zur Schwimmblase und Seitenlinie. no. 60, pp. 902-912.
Zeitschr. Morph. Okol. Tiere, vol. 31, WRIGHT, R. R.
pp. 371-410. 1884. The relationship between the air-bladder
1937. Anatomische Untersuchungen uiber die and auditory organ in Amiurus. Zool.
Seitenkanile der Sardine (Clupea pit- Anz., vol. 7, pp. 248-452.
chardus Walb.). Ibid., vol. 33, pp. 381- 1885. On the skull and auditory organs of
411. siluroid Hypophthalmus. Trans. Roy.
WOODWARD, A. S. Soc. Canada, vol. 3, pp. 107-118.
1901. Catalogue of the fossil fishes in the
 INDEX TO NAMES IN PROVISIONAL OUTLINE CLASSIFICATION
OF THE TELEOSTEAN FISHES
Abudefdufidae, 401 Amphiprionidae, 401
Abyssocottidae, 399 Amphiprioniformes, 399
Acanthoclinidae, 399 Amphisilidae, 398
Acanthocybiidae, 402 Anabantidae, 402
Acanthopsidae, 396 Anabantiformes, 399
Acanthopterygii, 398 Anabantoidei, 402
Acanthuridae, 402 Anablepidae, 397
Acanthuriformes, 399 Anablepsidae, 397
Acanthuroidei, 402 Anacanthidae, 403
Aceratiidae, 397 Anacanthini, 397
Acestrorhynchidae, 395 Anarhichadidae, 401
Achiridae, 403 Anarrhichadidae, 401
Acinaceidae, 402 Anarrhichthyidae, 401
Acronuridae, 402 Anguillichthyidae, 393
Acropomatidae, 400 Anguillidae, 393
Acropomidae, 400 Anguilliformes, 393
Acrotidae, 402 Anguillimorphi, 393
Adiposiidae, 396 Anguilloidei, 393
Adrianichthyidae, 397 Anisochromidae, 399
Aeschynichthyidae, 397 Anodontidae, 395
Ageneiosidae, 396 Anomalopidae, 398
Agonidae, 399 Anoplogasteridae, 398
Agriopidae, 399 Anoplopomatidae, 399
Akysidae, 396 Anoplopomidae, 399
Alabetidae, 398 Anostomidae, 395
Alabetoidei, 398 Anotopteridae, 395
Alabidae, 398 Antennariidae, 397
Alabiformes, 398 Antennarioidei, 397
Albulidae, 393 Anthiidae, 399
Albuloidei, 393 Antigoniidae, 398
Alepidosauridae, 395 Aoteidae, 393
Alepisauridae, 395 Aoteridae, 393
Alepocephalidae, 394 Aphareidae, 400
Alepocephaloidei, 394 Aphredoderidae, 396
Aleuteridae, 403 Aphredoderoidei, 396
Allotriognathi, 398 Aphyonidae, 397
Aluteridae, 403 Aploactidae, 399
Ambassidae, 399 Aploactinidae, 399
Ambassiidae, 399 Aplochitonidae, 394
Ambiotocidae, 400 Aplodactylidae, 401
Amblycepidae, 396 Apocrypteidae, 402
Amblycipitidae, 396 Apodes, 393
Amblyopidae, 402 Apogonichthyidae, 400
Amblyopsidae, 396 Apogonidae, 400
Amblyopsiformes, 396 Apolectidae, 400
Amblyopsoidei, 396 Apteronotidae, 396
Ameiuridae, 396 Aracanidae, 403
"Amiidae," 400 Arapaimidae, 394
Amiuridae, 396 Argentinidae, 394
Ammodytidae, 402 Argentinoidei, 394
Ammodytoidei, 402 Argidae, 396
Amphacanthidae, 402 Ariidae, 396
Amphiliidae, 396 Arripidae, 400
Amphipnoidae, 399 Arripididae, 400
445
446 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VO L. 131
Ascelichthyidae, 399 Bembropsidae, 401
Aspidophoridae, 399 Benthophilidae, 402
Aspidophoroididae, 399 Benthosauridae, 395
Aspredinidae, 396 Berycidae, 398
Astroblepidae, 396 Beryciformes, 398, 399
Astronesthidae, 394 Berycoidei, 398, 399
Astroscopidae, 401 Berycomorphi, 398
Ateleopidae, 395 Bivibranchiidae, 395
Ateleopiformes, 395 Bleekeridae, 402
Ateleopodidae, 395 Bleekeriidae, 402
Ateleopodoidei, 395 Blenniidae, 401
Atherinidae, 397 Blenniiformes, 399
Atheriniformes, 397 Blennioidei, 401
Atherinoidei, 397 Blepisiidae, 399
Atherinomorpha, 397 Blepsiidae, 399
Atopoclinidae, 401 Bodianidae, 401
Auchenipteridae, 396 Bostockiidae, 399
Aulopidae, 395 Bothidae, 402
Aulopodidae, 395 Bovichthyidae, 401
Aulorhynchidae, 398 Bovichtidae, 401
Aulostomi, 398 Bovictidae, 401
Aulostomidae, 398 Branchionichthyidae, 397
Aulostomiformes, 398 Bramidae, 400
Aulostomoidei, 398 Branchiostegidae, 400
Avocettinidae, 393 Bregmacerotidae, 397
Avocettinopsidae, 393 Brotulidae, 397
Brotulophidae, 397
Bagariidae, 396 Bunocephalidae, 396
Bagreidae, 396
Bagridae, 396 Caesiodidae, 400
Balistidae, 403 Caesionidae, 400
Balistoidei, 403 Callichthyidae, 396
Banjosidae, 400 Callionymidae, 402
Barbourisidae, 395 Callionymoidei, 402
Barbourisiidae, 395 Callipterygidae, 401
Bathyaploactidae, 399 Callophysidae, 396
Bathyclupeidae, 400 Callyodontidae, 401
Bathyclupeiformes, 399 Canthigasteridae, 403
Bathydraconidae, 401 Caproidae, 398
Bathylaconidae, 394 Caprophonidae, 398
Bathylaconoidei, 394 Caracanthidae, 399
Bathylagidae, 394 Carangidae, 400
Bathymasteridae, 401 Carangiformes, 399
Bathypteridae, 395 Carapidae, 397
Bathypteroidae, 395 Caristiidae, 400
Bathythrissidae, 393 Cataphracti, 399
Bathysauridae, 395 Catostomidae, 396
Batrachidae, 396 Caulolepidae, 398
Batrachoididae, 396 Caulophrynidae, 397
Batrachoidiformes, 396 Cebedichthyidae, 401
Bedotiidae, 397 Centracanthidae, 400
Belonidae, 397 Centracantidae, 400
Beloniformes, 397 Centrarchidae, 400
Belontidae, 402 Centriscidae, 398
Belontiidae, 402 Centrolophidae, 402
Bembradidae, 399 Centrophrynidae, 397
Bembridae, 399 Centropomidae, 399
Bembropidae, 401 Cephalacanthidae, 399
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 4L47
Cephalopholidae, 399 Clariidae, 396
Cepolidae, 401 Clinidae, 401
Ceratiidae, 397 Clupanodontidae, 394
Ceratioidei, 397 Clupeidae, 394
Cerdalidae, 402 Clupeiformes, 393, 394, 395
Cetomimidae, 395 Clupeoidei, 394
Cetomimiformes, 395 Clupeomorpha, 394
Cetomimoidei, 395 Clupidae, 394
Cetopsidae, 396 Clupisudidae, 394
Chacidae, 396 Cobitidae, 396
Chaenichthyidae, 401 Colocongridae, 393
Chaenopsidae, 401 Colomesidae, 403
Chaetodipteridae, 400 Comephoridae, 399
Chaetodontidae, 400 Congeridae, 393
Champsodontidae, 401 Congiopodidae, 399
Chandidae, 399 Congiopodoidei, 399
Chanidae, 395 Congothrissidae, 394
Channichthyidae, 401 Congridae, 393
Channicthyidae, 401 Congrogadidae, 401
Channidae, 398 Coracinidae, 400
Channiformes, 398 Coregonidae, 394
Chanoidae, 395 Coridae, 401
Chanoidei, 395 Coryphaenidae, 400
Chanoiformes, 395 Coryphaeniformes, 399
Characidae, 395 Coryphaenoididae, 397
Characinidae, 395 Cottidae, 399
Characodontidae, 397 Cottocomephoridae, 399
Characoidei, 395 Cottoidei, 399
Chaudhuriidae, 402 Cottomorphi, 399
Chaudhuriiformes, 399 Cottunculidae, 399
Chauliodidae, 394 Craniomi, 399
Chauliodontidae, 394 Cranoglanididae, 396
Chaunacidae, 397 Creagrutidae, 395
Cheilobranchidae, 398 Creediidae, 401
Cheilodactylidae, 401 Crenuchidae, 395
Cheilodipteridae, 400 Cristidae, 400
Cheimarrhichthyidae, 401 Cromeriidae, 395
Chiasmodontidae, 401 Cryptacanthidae, 401
Chilobranchidae, 398 Cryptacanthodidae, 401
Chilodontidae, 395 Ctenolabridae, 401
Chilorhinidae, 393 Ctenoluciidae, 395
Chimarrichthyidae, 401 Ctenothrissiformes, 395
Chiridae, 399 Curimatidae, 395
Chirocentridae, 394 Cybiidae, 402
Chirolophidae, 401 Cyclogasteridae, 399
Chironemidae, 401 Cyclolabridae, 401
Chlopsidae, 393 Cyclopidae, 396
Chlorophthalmidae, 395 Cyclopiidae, 396
Chonarhinidae, 403 Cyclopteridae, 399
Chondrobrachii, 395 Cyemidae, 393
Chromidae, 401 Cynodontidae, 395
Chromides, 399 Cynoglossidae, 403
Chromileptidae, 399 Cyphosidae, 400
Cibiidae, 402 Cyprinidae, 396
Cichlidae, 401 Cypriniformes, 395, 396
Cirrhitidae, 401 Cyprinodontes, 396
Citharidae, 402 Cyprinodontidae, 397
Citharinidae, 396 Cyprinodontiformes, 396, 397
448 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
Cyprinodontoidei, 397 Elopiformes, 393
Cyprinoidei, 396 Elopoidei, 393
Cyttidae, 398 Elopomorpha, 393
Cyttopsidae, 398 Elopsidae, 393
Embiotocidae, 400
Dactylopteridae, 399 Embiotocomorphi, 399
Dactylopteriformes, 399 Emblemariidae, 401
Dactyloscopidae, 401 Emmelichthyidae, 400
Dalliidae, 394 Empetrichthyidae, 397
Denticidae, 400 Engraulidae, 394
Denticipitidae, 394 Engraulididae, 394
Denticipitoidei, 394 Enoplosidae, 400
Derepodichthyidae, 397 Epigonidae, 400
Derichthyidae, 393 Ephippidae, 400
Diademichthyidae, 397 Epinephelidae, 399
Dianidae, 402 Equulidae, 400
Diceratiidae, 397 Eretmophoridae, 397
Dichistiidae, 400 Ereuniidae, 399
Dinolestidae, 400 Erilepidae, 399
Diodontidae, 403 Erythrichthyidae, 400
Diodontomorphi, 403 Erythrinidae, 395
Diplomystidae, 396 Erythroclidae, 400
Diploprionidae, 399 Esocesidae, 397
Dipterygonotidae, 400 Esocidae, 394, 397
Diretmidae, 398 Esocoidei, 394
Discocephali, 399 Etelidae, 400
Disparichthyidae, 397 Etheostomatidae, 400
Distichodontidae, 395 Etheostomidae, 400
Ditremidae, 400 Eucentrarchidae, 400
Doiichthyidae, 396 Eucinostomidae, 400
Dolichopterygidae, 394 Eurypharyngidae, 393
Doliichthyidae, 402 Eutaeniophoridae, 395
Doradidae, 396 Eutelichthyidae, 399
Dorosomatidae, 394 Eventognathi, 395
Dorosomidae, 394 Evermannellidae, 395
Draconettidae, 402 Evolantiidae, 397
Drepanichthyidae, 400 Exocoetidae, 397
Drepanidae, 400 Exocoetoidei, 397
Duleidae, 400
Dussumeriidae, 394 Fierasferidae, 397
Dussumieridae, 394 Fistularidae, 398
Dussumieriidae, 394 Fistulariidae, 398
Dysominidae, 393 Fitzroyidae, 397
Dysommidae, 393 Fitzroyiidae, 397
Dysomminidae, 393 Flutidae, 399
Formiidae, 400
Echelidae, 393 Formionidae, 400
Echeneidae, 400 Fundulidae, 397
Echeneididae, 400
Echeneiformes, 399 Gadidae, 397
Echidnidae, 393 Gadiformes, 397
Elacatidae, 400 Gadoidei, 397
Elassomatidae, 400 Gadomorphi, 397
Elassomidae, 400 Gadopseiformes, 399
Electrophoridae, 396 Gadopsidae, 401
Eleotridae, 402 Gaidropsaridae, 397
Elephenoridae, 400 Galaxidae, 394
Elopidae, 393 Galaxiidae, 394
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 449
Galaxiiformes, 394 Gyrinocheilidae, 396
Galaxioidei, 394
Gambusiformes, 397 Haemulidae, 400
Gasterochismidae, 402 Haemulonidae, 400
Gasteropelecidae, 395 Halidesmidae, 401
Gasterosteidae, 398 Haliophidae, 401
Gasterosteiformes, 398 Halosauridae, 393
Gasterosteoidei, 398 Halosauriformes, 393
Gastromyzonidae, 396 Haplochitonidae, 394
Gastromyzontidae, 396 Haplodactylidae, 401
Gastropelecidae, 395 Haplodoci, 396
Gaterinidae, 400 Haplomi, 394
Gavialicipitidae, 393 Harpadontidae, 395
Gelididae, 401 Harpagiferidae, 401
Gempylidae, 402 Harpidae, 401
Gerridae, 400 Harpodontidae, 395
Gibberichthyidae, 398 Harpuridae, 402
Gigantactinidae, 397 Helogeneidae, 396
Giganturidae, 395 Helostomatidae, 402
Giganturiformes, 395 Helostomidae, 402
Giganturoidei, 395 Hemerocoetidae, 401
Girellidae, 400 Hemiodidae, 395
Glandulocaudidae, 395 Hemiodontidae, 395
Glanencheli, 395 Hemiramphidae, 397
Glaucosomidae, 400 Hemirhamphidae, 397
Glyphiodontidae, 401 Hemitripteridae, 399
Gnathanacanthidae, 399 Henichthyidae, 400
Gobiesocidae, 397 Henicichthyidae, 400
Gobiesociformes, 397 Hepatidae, 402
Gobiesocomorphi, 397 Hepsetidae, 395
Gobiidae, 402 Heterenchelidae, 393
Gobiiformes, 399 Heterenchelyidae, 393
Gobiobotidae, 396 Heterocongridae, 393
Gobiodidae, 402 Heterognathi, 395
Gobioidea, 399 Heteromi, 393
Gobioidei, 402 Heteromyridae, 393
Gobiomoridae, 402 Heteropneustidae, 396
Gonorhynchidae, 395 Heterosomata, 402
Gonorynchidae, 395 Heterotidae, 394
Gonorynchiformes, 395 Hexagrammidae, 399
Gonorynchoidei, 395 Hexagrammoidei, 399
Gonostomatidae, 394 Himantolophidae, 397
Gonostomidae, 394 Hiodontidae, 394
Goodeidae, 397 Hippocampidae, 398
Grammicolepidae, 398 Hippoglossidae, 402
Grammicolepididae, 398 Histiophoridae, 402
Grammidae, 399 Histiopteridae, 400
Grammistidae, 399 Holconoti, 399
Grasseichthyidae, 395 Holconotidae, 400
Gregoryinidae, 400 Holocenthridae, 398
Grystidae, 400 Holocentridae, 398
Gunnellichthyidae, 402 Hologenidae, 396
Gymnapogonidae, 400 Homalopteridae, 396
Gymnarchidae, 394 Hoplegnathidae, 400
Gymnodontidae, 403 Hoplichthyidae, 399
Gymnonoti, 395 Hoplichthyoidei, 399
Gymnotidae, 396 Hoplopagridae, 400
Gymnotoidei, 396 Hoplopterygidae, 398
450 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VO L. 131
Horaichthyidae, 397 Latidae, 399
Hyodontidae, 394 Latilidae, 400
Hypophthalmidae, 396 Latridae, 401
Hypoplectrodidae, 399 Latrididae, 401
Hypoptychidae, 402 Lebiasinidae, 395
Hypostomidae, 396 Leiognathidae, 400
Hypostomides, 399 Lemnisomidae, 402
Hypsaeidae, 396 Lepidoglanidae, 396
Hypsocidae, 396 Lepidopidae, 402
Hysterocarpidae, 400 Lepidotidae, 400
Lepodidae, 400
Icelidae, 399 Leptobramidae, 400
Ichthyboridae, 396 Leptocephalidae, 393
Ichthyoboridae, 396 Leptoscopidae, 401
Icichthyidae, 402 Lethrinidae, 400
Icosteidae, 402 Limnichthidae, 401
Icosteiformes, 399 Limnichthyidae, 401
Icosteoidei, 402 Linophrynidae, 397
Ictaluridae, 396 Liognathidae, 400
Idiacanthidae, 394 Liparidae, 399
Igborichthyidae, 394 Lipariidae, 399
Ilarchidae, 400 Liparopidae, 399
Ilyophidae, 393 Lipogenidae, 394
Indostomidae, 398 Lipogenyidae, 394
Inermiidae, 400 Lobotidae, 400
Iniomi, 394 Lophiidae, 397
Ipnopidae, 395 Lophiiformes, 397
Isonidae, 398 Lophioidei, 397
Isospondyli, 393, 394 395 Lophobranchii, 398
Istiophoridae, 402 Lophotidae, 398
Loricariidae, 396
Jenynsiidae, 397 Loricati, 399
Jordaniidae, 399 Luciidae, 394
Jugulares, 396, 399 Luciocephalidae, 402
Juvenellidae, 400 Luciocephaloidei, 402
Kasidoridae, 395 Lumpenidae, 401
Kasidoroidae, 395 Luthianidae, 400
Katsuwonidae, 402 Lutianidae, 400
Kneriidae, 395 Lutjanidae, 400
Korsogasteridae, 398 Luvaridae, 402
Kraemeriidae, 402 Lycodapodidae, 397
Kuhliidae, 400 Lycodidae, 397
Kurtidae, 402 Lyconidae, 397
Kurtoidei, 402 Lyomeri, 393
Kyphosidae, 400 Lyopomi, 393
Labracoglossidae, 400 Maccullochellidae, 399
Labridae, 401 Macquariidae, 399
Labroidei, 401 Macristiidae, 395
Labyrinthici, 398, 399 Macrocephanchelidae, 393
Labyrinthicidae, 402 Macrocephenchelyidae, 393
Lactariidae, 400 Macropinnidae, 394
Laevoceratiidae, 397 Macroramphosidae, 398
Lagocephalidae, 403 Macrorhamphosidae, 398
Lampridae, 398 Macrouridae, 397
Lamprididae, 398 Macrouroidei, 397
Lampridiformes, 398 Macrouroididae, 397
Lampridoidei, 398 Macruridae, 397
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 451
Macruriformes, 397 Mullidae, 400
Macrurocyttidae, 398 Muraenesocidae, 393
Maenidae, 400 Muraenichthyidae, 393
Makairidae, 402 Muraenidae, 393
Malacanthidae, 400 Muraenolepidae, 397
Malacichthyes, 399 Muraenolepididae, 397
Malacosteidae, 394 Muraenolepoidei, 397
Malapteruridae, 396 Myctophidae, 395
Malopteruridae, 396 Myctophiformes, 394
Maltheidae, 397 Myctophoidei, 395
Malthidae, 397 Myridae, 393
Marukawichthyidae, 399 Myrocongridae, 393
Mastacembelidae, 402 Myrophidae, 393
Mastacembeliformes, 399 Mystidae, 396
Mastacembeloidei, 402
Maurolicidae, 394 Nandidae, 400
Medidae, 396 Nannatherinidae, 400
Megalopidae, 393 Nannopercidae, 400
Melamphaeidae, 398 Nannostomidae, 395
Melamphaidae, 398 Nasidae, 402
Melamphasidae, 398 Neenchelyidae, 393
Melanocetidae, 397 Nematistiidae, 400
Melanostomiatidae, 394 Nematognathi, 396
Melanotaeniidae, 397 Nemichthyidae, 393
Menidae, 400 Nemipteridae, 400
Merlucciidae, 397 Nemophididae, 401
Merolepidae, 400 Neoatherinidae, 397
Microcyprini, 396, 397 Neoceratiidae, 397
Microdesmidae, 402 Neodacidae, 401
Micropteridae, 400 Neodaciidae, 401
Microstomatidae, 394 Neolabridae, 401
Microstomidae, 394 Neolethrinidae, 400
Milyeringidae, 402 Neophrynichthyidae, 399
Mirapinnatidae, 395 Neoscopelidae, 395
Mirapinnatoidei, 395 Neostethidae, 398
Mirapinnidae, 395 Nessorhamphidae, 393
Mochochidae, 396 Nettastomatidae, 393
Mochocidae, 396 Nettastomidae, 393
Mochockidae, 396 Niphonidae, 399
Mochokidae, 396 Nomeidae, 402
Molidae, 403 Normanichthyidae, 399
Monacanthidae, 403 Notacanthidae, 394
Monocentridae, 398 Notacanthiformes, 393
Monodactylidae, 400 Notograptidae, 401
Monognathidae, 393 Notopteridae, 394
Monognathiformes, 393 Notopteroidei, 394
Monopteridae, 399 Notosudidae, 395
Monotaxidae, 400 Nototheniidae, 401
Moridae, 397 Notothenioidei, 401
Moringuidae, 393 Novumbridae, 394
Mormyridae, 394
Mormyriformes, 394 Odacidae, 401
Moronidae, 399 Odontostomidae, 395
Mugilidae, 401 Ogcocephalidae, 397
Mugiliformes, 397, 399 Oligoridae, 399
Mugiloidei, 401 Olyridae, 396
Mugiloididae, 401 Omosudidae, 395
Mugilomorphi, 397 Onchocephalidae, 397
452 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 131
Oncocephalidae, 397 Parapercidae, 401
Oneirodidae, 397 Parascorpidae, 400
Ophicephalidae, 398 Parazenidae, 398
Ophichthidae, 393 Pareioplitae, 399
Ophichthyidae, 393 Parodontidae, 395
Ophiclinidae, 401 Parophiocephalidae, 398
Ophidiidae, 397 Pataecidae, 399
Ophidioidei, 397 Pediculati, 396, 397
Ophiocephalidae, 398 Pegasidae, 399
Ophiocephaliformes, 398 Pegasiformes, 399
Ophioclinidae, 401 Pempheridae, 400
Ophiodontidae, 399 Pentacerotidae, 400
Opisthocentridae, 402 Pentapodidae, 400
Opisthognathidae, 401 Percesoces, 397, 399
Opisthomi, 399 Percichthyidae, 399
Opisthoproctidae, 394 Percidae, 400
Oplegnathidae, 400 Perciformes, 396, 397, 399
Oplichthyidae, 399 Perciliidae, 399
Oreosomatidae, 398 Percoidei, 399
Orestiidae, 397 Percomorphi, 399
Orthagoriscidae, 403 Percophidae, 401
Oryziatidae, 397 Percophididae, 401
Osmeridae, 394 Percopsidae, 396
Osphromenidae, 402 Percopsiformes, 396
Osphronemidae, 402 Percopsoidei, 396
Ostariophysi, 395 Percopsomorphi, 396
Osteoglossidae, 394 Periophthalmidae, 402
Osteoglossiformes, 394 Peristediidae, 399
Osteoglossoidei, 394 Peristediontidae, 399
Osteoglossomorpha, 394 Peronedyidae, 401
Ostorhinchidae, 400 Peronedysidae, 401
Ostraciidae, 403 Petalichthyidae, 397
Ostraciontidae, 403 Phallostethidae, 398
Ostracoberycidae, 399 Phallostethiformes, 397
Otolithidae, 400 Pharopterycidae, 399
Ovoididae, 403 Pharyngognathi, 399
Owstoniidae, 401 Pholidae, 402
Oxudercidae, 401 Pholidichthyidae, 402
Oxylabracidae, 399 Pholididae, 402
Oxylebiidae, 399 Photocorynidae, 397
Oxyporhamphidae, 397 Phractolaemidae, 395
Pimelepteridae, 400
Pampidae, 402 Pimelodidae, 396
Pangasiidae, 396 Pinguipedidae, 401
Pantodontidae, 394 Plagyodontidae, 395
Parabembridae, 399 Planidae, 402
Paracanthopterygii, 396 Platacidae, 400
Paracentropristidae, 399 Platycephalidae, 399
Paraclinidae, 401 Platycephaloidei, 399
Paradichthyidae, 400 Platyproctidae, 394
Paradicichthyidae, 400 Platypteridae, 402
Paragalaxiidae, 394 Platytroctidae, 394
Paragobioididae, 402 Plecoglossidae, 394
Paralepidae, 395 Plecostei, 399
Paralepididae, 395 Plectognathi, 403
Paralichthodidae, 402 Plectorhynchidae, 400
Paralichthyidae, 402 Plectospondyli, 395, 396
Parapercichthyidae, 401 Plectroplitidae, 399
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 453
Plectropomidae, 399 Regalecidae, 398
Plesiopidae, 399 Retropinnidae, 394
Pleuronectidae, 402 Rhachycentridae, 400
Pleuronectiformes, 402 Rhamphichthyidae, 396
Pleuronectoidei, 402 Rhamphocottidae, 399
Pleuroscopidae, 401 Rhamphosidae, 398
Plotosidae, 396 Rhamphosiformes, 398
Podatelidae, 395 Rhegmatidae, 399
Poeciliidae, 397 Rhegnopteri, 399
Polyacanthidae, 402 Rhinoprenidae, 400
Polycentridae, 400 Rhodichthyidae, 399
Polymixiidae, 398 Rhombosoleidae, 402
Polymixioidei, 398 Rhyacichthyidae, 402
Polynemidae, 401 Rhynchobdellidae, 402
Polynemiformes, 399 Rondeletiidae, 395
Polynemoidei, 401 Rosauridae, 395
Pomacanthidae, 400 Runulidae, 401
Pomacentridae, 401 Ruvettidae, 402
Pomadasidae, 400 Rypticidae, 399
Pomadasyidae, 400
Pomatomidae, 400 Saccobranchidae, 396
Porcidae, 396 Saccopharyngidae, 393
Premnidae, 401 Saccopharyngiformes, 393
Prenidae, 400 Saccopharyngoidei, 393
Priacanthidae, 400 Salangidae, 394
Pristigasteridae, 394 Salariidae, 401
Pristipomatidae, 400 Salmonidae, 394
Pristipomidae, 400 Salmoniformes, 394
Pristolepidae, 400 Salmonoidei, 394
Prochilodontidae, 395 Salmopercae, 396
Protacanthopterygii, 394 Samaridae, 402
Prototroctidae, 394 Sardidae, 402
Psammichthyidae, 402 Sauridae, 395
Psenidae, 402 Sauromuraenesocidae, 393
Psettidae, 400 Scarichthyidae, 401
Psettodidae, 402 Scaridae, 401
Psettodoidei, 402 Scatophagidae, 400
Pseudaphritidae, 401 Schilbeidae, 396
Pseudochromidae, 399 Schindleriidae, 402
Pseudogrammidae, 399 Schindlerioidei, 402
Pseudomugilidae, 397 Sciaenidae, 400
Pseudopimelodidae, 396 Sclerogeni, 399
Pseudoplesiopidae, 399 Sclerogenidae, 398, 399
Psilocephalidae, 403 Scleroparei, 398, 399
Psilorhynchidae, 396 Scolopsidae, 400
Psychrolutidae, 399 Scomberesocidae, 397
Pteraclidae, 400 Scomberidae, 402
Pteraclididae, 400 Scombermoridae, 402
Pteropsaridae, 401 Scombresocidae, 397
Pterothrissidae, 393 Scombridae, 402
Ptilichthyidae, 401 Scombriformes, 399
Pygidiidae, 396 Scombroidei, 402
Pyramodontidae, 397 Scombropidae, 400
Scomibropsidae, 400
Rachycentridae, 400 Scopelarchidae, 395
Rainfordiidae, 399 Scopelidae, 395
Ranicipitidae, 397 Scopeliformes, 394, 395
Ratabouridae, 393 Scopelosauridae, 395
454 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VO L. 131
Scophthalmidae, 402 Stromateidae, 402
Scorpaenichthyidae, 399 Stromateoidei, 402
Scorpaenidae, 399 Stylephoridae, 398
Scorpaeniformes, 399 Stylephoroidei, 398
Scorpaenoidei, 399 Stylophoridae, 398
Scorpidae, 400 Stylophthalmidae, 394
Scorpididae, 400 Stylophthalrnoidae, 394
Scyphophori, 394 Sudidae, 395
Scytalinidae, 402 Symbranchia, 398
Scytaliscidae, 402 Symbranchidae, 399
Searsidae, 394 Symbranchiformes, 398
Searsiidae, 394 Symbranchii, 398
Selenichthyes, 398 Synanceidae, 399
Seriolidae, 400 Synancejidae, 399
Serranidae, 399 Synaphobranchidae, 393
Serrasalmidae, 395 Synapturidae, 403
Serrivomeridae, 393 Synbranchidae, 399
Sicydiaphiidae, 402 Synbranchiformes, 398
Siganidae, 402 Synbranchoidei, 398
Sillaginidae, 400 Synchiridae, 399
Siluridae, 396 Synentognathi, 397
Siluriformes, 396 Syngnathidae, 398
Siluroidiformes, 396 Syngnathiformes, 398
Simenchelidae, 393 Syngnathoidei, 398
Simenchelyidae, 393 Synodidae, 395,396
Siphonognathidae, 401 Synodontidae, 395
Siphostomidae, 398
Sisoridae, 396 Tachysuridae, 396
Soleidae, 403 Taenioidae, 402
Solenichthyes, 398 Taenioididae, 402
Solenichthyidae, 398 Taeniophoridae, 395
Solenostomatichthyidae, 398 Teraponidae, 400
Solenostomatidae, 398 Terapontidae, 400
Solenostomidae, 398 Tesapontidae, 400
Soleoidei, 403 Tetragonopteridae, 395
Sorosichthyidae, 398 Tetragonuridae, 402
Sparidae, 400 Tetraodontidae, 403
Sparisomidae, 401 Tetraodontiformes, 403
Sphoeroididae, 403 Tetraodontoidei, 403
Sphyraenidae, 401 Tetrapturidae, 402
Sphyraenoidei, 401 Tetrarogidae, 399
Spicaridae, 400 Tetrodontidae, 403
Squamipenes, 399 Teuthidae, 402
Steinegeriidae, 400 Teuthididae, 402
Stephanoberycidae, 398 Theraponidae, 400
Stephanoberyciformes, 395, 398 Theutyidae, 402
Stephanoberycoidei, 398 Thoracostei, 398
Sternarchidae, 396 Thunnidae, 402
Sternoptychidae, 394 Thunniformes, 399
Sternoptychiidae. 394 Thymallidae, 394
Sternopygidae, 396 Todaridae, 393
Stichaeidae, 401 Tomeuridae, 397
Stichariidae, 401 Torpedinidae, 396
Stigmatonotidae, 399 Toxotidae, 400
Stilbiscidae, 393 Trachichthyidae, 398
Stolephoridae, 394 Trachinidae, 401
Stomiatidae, 394 Trachiniformes, 399
Stomiatoidei, 394 Trachinoidei, 401
1966 GREENWOOD, ROSEN, WEITZMAN, MYERS: FISHES 455
Trachipteridae, 398 Xenarchi, 396
Trachipteroidei, 398 Xenichthyidae, 400
Trachyberycidae, 400 Xenoberyces, 398, 399
Trachycorystidae, 396 Xenocephalidae, 401
Trachypteridae, 398 Xenocongridae, 393
Triacanthidae, 403 Xenomi, 394
Triacanthodidae, 403 Xenophthalmichthyidae, 394
Trichiuridae, 402 Xenopoclinidae, 401
Trichodontidae, 401 Xenopteridae, 403
Trichomycteridae, 396 Xenopterygii, 397
Trichonotidae, 401 Xiphasiidae, 401
Triglidae, 399 Xiphidiontidae, 401
Trinectidae, 403 Xiphiidae, 402
Triodontidae, 403 Xiphisteridae, 401
Tripterophycidae, 397 Xiphostomatidae, 395
Tripterygiidae, 401 Xiphostomidae, 395
Tripterygiontidae, 401 Xystaemidae, 400
Triuridae, 403
Tropidichthyidae, 403
Trypauchenidae, 402 Zanclidae, 402
Tylosuridae, 397 Zaniolepidae, 399
Zaniolepididae, 399
Umbridae, 394 Zanteclidae, 397
Uranoscopidae, 401 Zaproridae, 402
Zeidae, 398
Zeiformes, 398
Veliferidae, 398 Zenidae, 398
Veliferoidei, 398 Zeniontidae, 398
Verilidae, 400 Zeoidei, 398
Winteridae, 394
Zeomorphi, 398
Winteriidae, 394
Zoarcidae, 397
Zoarcoidei, 397