A Phylogenetic Study of the
Neotropical Characiform Family
Curimatidae (Pisces: Ostariophysi)
RICHARD P. VARI
Ifc
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY ? NUMBER 471
SERIES PUBLICATIONS OF THE SMITHSONIAN INSTITUTION
Emphasis upon publication as a means of "diffusing knowledge" was expressed by the first
Secretary of the Smithsonian. In his formal plan for the Institution, Joseph Henry outlined a
program that included the following statement: "It is proposed to publish a series of reports,
giving an account of the new discoveries in science, and of the changes made from year to year
in all branches of knowledge." This theme of basic research has been adhered to through the
years by thousands of titles issued in series publications under the Smithsonian imprint,
commencing with Smithsonian Contributions to Knowledge in 1848 and continuing with the
following active series:
Smithsonian Contributions to Anthropology
Smithsonian Contributions to Astrophysics
Smithsonian Contributions to Botany
Smithsonian Contributions to the Earth Sciences
Smithsonian Contributions to the Marine Sciences
Smithsonian Contributions to Paleobiology
Smithsonian Contributions to Zoology
Smithsonian Folklife Studies
Smithsonian Studies in Air and Space
Smithsonian Studies in History and Technology
In these series, the Institution publishes small papers and full-scale monographs that report
the research and collections of its various museums and bureaux or of professional colleagues
in the world of science and scholarship. The publications are distributed by mailing lists to
libraries, universities, and similar institutions throughout the world.
Papers or monographs submitted for series publication are received by the Smithsonian
Institution Press, subject to its own review for format and style, only through departments of the
various Smithsonian museums or bureaux, where the manuscripts are given substantive review.
Press requirements for manuscript and art preparation are outlined on the inside back cover.
Robert McC. Adams
Secretary
Smithsonian Institution
S M I T H S O N I A N C O N T R I B U T I O N S T O Z O O L O G Y ? N U M B E R 4 7 1
A Phylogenetic Study of the
Neotropical Characiform Family
Curimatidae (Pisces: Ostariophysi)
Richard P. Vari
SMITHSONIAN INSTITUTION PRESS
Washington, D.C.
1989
A B S T R A C T
Vari, Richard P. A Phylogenetic Study of the Neotropical Characiform Family Curimatidae
(Pisces: Ostariophysi). Smithsonian Contributions to Zoology, number 471, 71 pages, 45
figures, 3 tables, 1989.?Osteology and soft anatomy of the species of Curimatidae were studied
to examine the hypothesis that the family is monophyletic, and to advance a hypothesis of
phylogenetic relationships within the family.
Features of the branchial apparatus, buccopharyngeal complex, hyoid arch, jaws, palatine
arch, and neurocranium corroborate the monophyly of the Curimatidae. Additional characters
in a subset of these systems are congruent with the hypothesis of a sister group relationship
between the Curimatidae and Prochilodontidae (Vari, 1983). Shared derived characters of those
systems and the antorbital, infraorbitals, opercular apparatus, ligamentum primordiale, hypural
complex, caudal-fin rays, proximal pterygiophores of the dorsal fin, laterosensory canal systems
of the head and body, and squamation define phylogenetic subunits of the Curimatidae,
supporting a phylogenetic hypothesis of four sequential dichotomies and a terminal
polychotomy.
I recognize eight genera in the Curimatidae. Curimatopsis, Potamorhina, Curimata,
Psectrogaster, Steindachnerina, Pseudocurimata, and Curimatella are each hypothesized as
monophyletic on the basis of additional shared derived characters. No unique derived characters
were found to define the eighth recognized genus, Cyphocharax.
Curimatichthys Fernandez-Yepez (1948) is a synonym of Curimatopsis. Potamorhina is
redefined to include Gasterotomus Eigenmann (1910), Gasterostomus Femandez-Yepez (1948),
Suprasinelepichthys Fernandez-Yepez (1948), and Potamorrhina Braga and Azpelicueta
(1983). Curimata Bosc (1817) has nine junior synonyms: Semitapicis Eigenmann and
Eigenmann (1889b); Peltapleura Fowler (1906); Acuticurimata Fowler (1941); Allenina
Fernandez-Yepez (1948); Lambepiedra Femandez-Yepez (1948); Bitricarinatra Fernandez-
Yepez (1948); Bondichthys Whitley (1953); Stupens Whitley (1954); and Semitapiscis Braga
and Azpilicueta (1983). Three genera are included in Psectrogaster. Pseudopsectrogaster
Fernandez-Yepez (1948), Hamatichthys Fernandez-Yepez (1948), and Semelcarinata Femandez-
Yepez (1948). Rivasella and Curimatorbis, both proposed by Femandez-Yepez (1948), are
synonyms of Steindachnerina Fowler (1906). Curimatella has three junior synonyms, all
described by Femandez-Yepez (1948): Apolinarella, Walbaunina, and Lepipinna. Cyphocharax
has four synonyms: Xyrocharax Fowler (1913b), Hemicurimata Myers (1929), Curimatoides
Fowler (1940), and Cruxentina Fernandez-Yepez (1948).
The hypothesis of intrafamilial phylogenetic relationships described herein is incongruent,
to differing degrees, with classificatory schemes proposed by Eigenmann and Eigenmann
(1889b), Eigenmann (1910), Fowler (1975), and Gery (1977b). Numerous differences exist at
all taxonomic levels between the scheme of relationships proposed in this study and the
"phylogenetic tree" put forward by Femandez-Yepez (1948). Additional hypothesized derived
characters described herein are in agreement with the reassignment of Anodus Spix, 1829 (in
Spix and Agassiz, 1829) and Eigenmannina Fowler (1906) from the Curimatidae to the
Hemiodontidae, first proposed by Roberts (1974) and congruent with results of Vari (1983).
Homoplasious characters are discussed in terms of their overall taxonomic distributions, and
relative to their innovative versus reductive natures. Reductive features involving the degree
of development of the laterosensory canal system were found to be convergent between
Curimatopsis and a subset of Cyphocharax. These paedomorphic features are evidently a
consequence of the relative size reduction of these species compared to proximate outgroups.
OFFICIAL PUBLICATION DATE is handstamped in a limited number of initial copies and is
recorded in the Institution's annual report, Smithsonian Year. SERIES COVER DESIGN: The coral
Montastrea cavernosa (Linnaeus).
Library of Congress Cataloging in Publication Data
Vari, Richard P.
A phylogcnetic study of the neotropical characiform family Curimatidae (Pisces: Ostariophysi).
(Smithsonian contributions to zoology ; no. 471)
Bibliography: p.
1. Curimatidae?Latin America?Classification. 2. Fishes?Latin America?Classification. I. Title. II. Series.
[QL638.C89V37] 1989 [597'.52] 88-15661
Contents
Page
Introduction 1
Acknowledgments 4
Systematic Procedures 4
Methods and Materials 5
Terminology 5
Material Examined 10
Abbreviations 11
Character Description and Analysis 12
Branchial Arches 12
Dorsal Portion of Branchial Arches 13
Fourth and Fifth Epibranchials 13
Fifth Upper Pharyngeal Tooth Plate 20
Fourth Upper Pharyngeal Tooth Plate and Fourth Infrapharyngobranchial. . . 21
Third Epibranchial 22
Third Infrapharyngobranchial 23
Second Branchial Arch 25
First Branchial Arch 25
Ventral Portion of Branchial Arches 26
First Basibranchial 26
Second Hypobranchial 27
Third Hypobranchial 27
Third Ceratobranchial 29
Fourth Basibranchial 29
Fourth Ceratobranchial 29
Fifth Ceratobranchial 31
Buccopharyngeal Complex 31
Hyoid Arch 33
Urohyal 33
Basihyal and Basihyal Tooth Plate 35
Branchiostegal Rays 36
Antorbital and Infraorbitals 36
Antorbital 37
First Infraorbital 37
Fourth and Fifth Infraorbitals 37
Sixth Infraorbital 39
Opercular Apparatus 40
Opercle and Subopercle 40
Interopercle 40
Ligamentum Primordiale 40
Upper and Lower Jaws 41
Premaxilla 41
Maxilla 42
Dentary 42
Palatine Arch 43
Palatine and Ethmo-palatine Cartilage 43
Ectopterygoid 44
iii
iv SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
Mesopterygoid 44
Metapterygoid 45
Supraneurals 47
Caudal Fin Complex 48
Laterosensory System of Body 49
Squamation 49
Synapomorphy List and Phylogenetic Reconstruction 49
Monophyly of the Curimatidae and Prochilodontidae Clade 51
Monophyly and Intrafamilial Relationships of the Family Curimatidae 52
Convergent Characters 59
Comparisons with Previous Classifications 63
Resumo 65
Literature Cited 67
A Phylogenetic Study of the
Neotropical Characiform Family
Curimatidae (Pisces: Ostariophysi)
Richard P. Vari
Introduction
The 100 or more species of the Neotropical characiform
family Curimatidae inhabit a broad range of freshwater
ecosystems from the still ox-bow lakes, slow-flowing streams,
and meandering rivers typical of lowland floodplains, to swift
tributaries and rapids of the Andean piedmont and the uplands
of the Guayanan and Brazilian shields. Curimatids are
components of the ichthyofauna of trans-Andean Pacific
drainages of Central and South America from southern Costa
Rica (Bussing, 1966:221) to northern Peru (Eigenmann,
1922:104; Ortega and Vari, 1986:11), and along that versant
they also inhabit the Caribbean drainages in northwestern
South America (most notably the Rio Magdalena system and
the Lago Maracaibo drainage basin). Curimatid species occur
in the vast portion of South America to the east of the Andes,
including the numerous rivers from the Rio Orinoco system
in Venezuela to south of Buenos Aires, Argentina.
Within this extensive geographic range, the most speciose
curimatid fauna is that of the Rio Amazonas basin, inhabited
by over 40 curimatid species representing all of the major
phyletic lineages of the family. The Rio Orinoco and Rio
Paraguay-Parana systems include over a dozen species of the
family within each of their ichthyofaunas. Less speciose
curimatid assemblages are found in Lago Maracaibo, the rivers
of the Atlantic slopes of the Guianas, the rivers of coastal
Brazil, Uruguay, and central Argentina, and in the Caribbean
and Pacific drainages of trans-Andean South and Central
America.
Richard P. Vari, Department of Vertebrate Zoology, National Museum
of Natural History, Smithsonian Institution, Washington, D.C. 20560.
Review Chairman: Gary R. Graves, National Museum of Natural
History, Smithsonian Institution. Reviewers: Gordon J. Howes, British
Museum (Natural History); Lynne R. Parenti, California Academy of
Sciences; Stanley H. Weitzman, National Museum of Natural History,
Smithsonian Institution.
The speciose nature of the Curimatidae is reflected in their
variability of overall body size and form. A nearly ten-fold
range in standard length (SL) is found between the largest
known adult males of Curimatopsis evelynae Gery (ca. 33
mm SL) and the largest reported Curimata mivartii Steindach-
ner (320 mm SL, Dahl, 1971:105). The range in body form
within the family is represented at one extreme by elongate
fusiform species such as Curimata ocellata Eigenmann and
Eigenmann (Figure 1) of the Amazon basin, which schools
with and mimics hemiodontids (Gery, 1977a). At the other
extreme, Psectrogaster abramoides (Kner) is a deep bodied,
laterally compressed species also from the Amazon (Figure 2).
The majority of curimatid species cluster around an intermedi-
ate, moderately compressed and slightly elongate body form
typified by Curimata incompta Vari (Figure 3), a species
endemic to the Rio Orinoco system.
Despite the poor representation of some members of the
family in systematic collections, many curimatid species occur
in large schools (Santos et al., 1985:28-29) that constitute a
significant proportion of fish biomass in riverine and lacustrine
ecosystems (Lowe-McConnell, 1975:109). In those localities,
they exploit the floculent organic material, microdetritus,
micro vegetation, and filamentous algae typical of most
Neotropical aquatic habitats (Carvalho, 1984; Nomura and
Hayashi, 1980; Nomura and Taveira, 1979). The populations
of numerous curimatid species are particularly concentrated
during spawning migrations such as the "piracema" of the Rio
Mogi-Guassu of southeastern Brazil (de Godoy, 1975:581-
603). These seasonal concentrations permit the ready exploita-
tion of those species by commercial and subsistence fisheries
throughout most of the distributional range of the family
(Mago-Leccia, 1970:31; Dahl, 1971:105; Rodriquez, 1973;
Lowe-McConnell, 1975:74; Saint-Paul and Bayley, 1979:112;
Smith, 1981:140; Goulding, 1981:60). Curimatids are also
important food items for larger, commercially important,
predatory fish species.
1
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
FIGURE 1.?Curimata ocellata, USNM 267973, 160.3 mm SL, Brazil, Amazonas, Rio Marauia.
FIGURE 2.?Cyphocharax abramoides, USNM 267952, 105.8 mm SL, Brazil, Para, Rio Xingu, Belo Monte.
FIGURE 3.?Curimata incompta, USNM 273308, 81.2 mm SL, Venezuela, Fstado Bolivar, still backwater off
Rio Orinoco at El Burro.
NUMBER 471
The diversity of curimatids, particularly the larger species,
have been long known to science. More than two centuries ago
both Linneaus (1758:312; 1766:514) and Gronovius (1763:123)
listed several nominal forms brought to Europe from
northeastern South America. Geographically and ecologically
more encompassing ichthyological collecting efforts in Neotro-
pical freshwaters since the middle of the eighteenth century
have brought curimatids of additional phyletic lineages to the
attention of systematists. This greater availability of curimatid
specimens from all regions of South America resulted in a
progressive increase of described species, with over 120
nominal species or subspecies proposed to date. Associated
with the increasingly complex species-level taxonomy were
the proposals of numerous genera and subgenera. Despite these
efforts, uncertainty continued to pervade curimatid taxonomy
at both the specific and supraspecific levels.
The multitude of species level problems within the
Curimatidae remained unsettled, partially as a consequence of
the absence of a broad overview of the group; most treatments
were limited to geographically or politically circumscribed
regions. These arbitrary limitations hindered elucidation of
species level problems, resolution of which required analysis
of geographic and taxonomic samples. The majority of authors
were further handicapped by absence of detail in many of the
original species descriptions, and by limited collections of
comparative materials. These factors, compounded by the large
number of nominal species of curimatids, resulted in uncer-
tainty as to the number of recognizable species of curimatids,
a poor understanding of the distinguishing characters of the
recognizable forms, and confusion as to the geographic
distributions of those taxa. Revisionary studies completed
(Vari, 1982a,b, 1984a,b, 1987; Vari and Gery, 1985; Vari and
Nijssen, 1986) and in progress focus on questions of specific
limits and geographic distributions. Such problems are not
pursued further in this paper.
Research on the second central problem in curimatid
taxonomy, the sequence of evolutionary events within the
family, has lagged behind analyses of species-level questions.
The relatively few studies that attempted, at least to some
degree, to address intra-familial phylogeny failed to clarify our
understanding of this core question. Those endeavors (e.g.,
Fernandez-Yepez, 1948) typically made little attempt to
critically examine the correlation, or lack thereof, between
recognized supraspecific taxa and hypothesized lineages within
the family. Rather, these studies usually proposed new genera
and subgenera based, either explicitly or by default, on degrees
of overall similarity or difference. In retrospect the weakness
of such a quasi-phenetic system is reflected both in the
erroneous specific and generic synonymizations that riddled
the taxonomy of the family, and in the inappropriate
assignment of species to supraspecific taxa. These factors led
to differing views as to the composition of generic and
suprageneric groupings within the Curimatidae.
The generic level subdivision of the family has proceeded
in an episodic fashion during the last two centuries.
Researchers prior to 1876 typically used Curimatus for the
vast majority of curimatids (Curimatus is equivalent to
Curimata Bosc, see Opinion 772, International Commission
of Zoological Nomenclature, 1966). The only exception was
the erroneous utilization of Anodus Spix for some members
of the family (e.g., Anodus latior Spix, 1829 (in Spix and
Agassiz, 1829) = Potamorhina latior, see Vari, 1984a). As
noted by Roberts (1974:429), Anodus is not available in the
Curimatidae, being properly applied to two species within the
Hemiodontidae (see "Synapomorphy List and Phylogenetic
Reconstruction"). Steindachner (1876:81) initiated the valid
generic level subdivision of the Curimatidae when he advanced
Curimatopsis as a subgenus of Curimata for the first species
of curimatid known to retain an incompletely pored laterosen-
sory canal as an adult. Cope (1878:675) continued this trend
towards increased generic diversity with his description of the
then monotypic genus Potamorhina for the distinctly modified
species Curimatus (Anodus) pristigaster previously described
by Steindachner (1876:73). The number of available nominal
genera and subgenera within the family underwent a slow but
progressive increase during the next seven decades in the
publications of the following authors: Eigenmann and Eigen-
mann, who proposed Psectrogaster (1889a:7), Curimatella
(1889b:415), and Semitapicis (1889b:417); Fowler, who
created Cyphocharax, Steindachnerina, Peltapleura (1906:297-
300), Xyrocharax (1913b:673), Curimatoides (1940:255), and
Acuticurimata (1941:162); Eigenmann (1910:422), who de-
scribed Gasterotomus; and Myers, (1929:620) who first used
Hemicurimata. Fernandez-Yepez (1948), in an outburst of
generic subdivision, more than doubled the number of nominal
genera and subgenera that were available prior to his study.
His newly proposed genera (Allenina, Apolinarella, Bitricari-
nata, Bondia, Camposella, Cruxentina, Curimatichthys, Curi-
matorbis, Hamatichthys, Lambepiedra, Lepipinna, Pseudocuri-
mata, Pseudopsectogaster, Rivasella, Semelcarinata, Suprasi-
nelepichthys, and Walbaunina) brought the number of generic
level names proposed for curimatids to thirty-one. Two of the
genera proposed by Fernandez-Yepez (1948), Bondia and
Camposella, were replaced with Bondichthys and Camposich-
thys by Whitley (1953:134) who pointed out that they were
already occupied in other groups. Camposichthys also proved
to be preoccupied, and Whitley (1954:30) put forward Stupens
as a substitute.
This trend towards generic level subdivision of the
Curimatidae, most notably by Fernandez-Yepez (1948), did
little to resolve questions of interrelationships within the family
at either specific or supraspecific levels. Ten of the twenty-
seven genera recognized by Fernandez-Yepez were monotypic
and could not serve as indicators of phylogenetic relationship.
The other more speciosc genera proposed or recognized by
Fernandez-Yepez were nearly always ambiguously defined,
with resultant uncertainties about generic limits and species
assignment, difficulties previously commented upon by Bohlke
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
(1958:106). Attempts to resolve the problems prevalent in the
classification proposed by Fernandez-Yepez resulted in the
subsequent shifting of species between nominal genera
(BOhlke, 1958:108; Ringuelet et al., 1967:198; Britski,
1969:203; Vari, 1982a: 12, 1984a: 11).
Questions of generic diagnosis and inclusiveness aside, we
nonetheless find that most authors who proposed classifications
for the family or its subunits did so without discussion of the
concepts, phenetic or phylogenetic, underlying their schemes.
That inexactitude inhibited evaluation of underlying evolution-
ary patterns hypothesized for the family as a whole or in part,
and obscured the degree to which the classifications attempted
to reflect a conceptual basis.
The single equivocal exception, the nomenclatural system
advanced by Fernandez-Yepez and reflected in his "phylo-
genetic tree" (1948, fig. 12), represented an amalgam of
phenetic and quasi-phylogenetic concepts based on a few
relatively labile characters analyzed neither in terms of
evolutionary polarity nor relative to their utility as indicators
of phylogenetic relationship. Indeed Fernandez-Yepez's (1948)
"phylogenetic tree" is a visual representation of his dichoto-
mous key to the genera of the family rather than a phylogenetic
scheme.
The generic level subdivision of the family by Fernandez-
Y6pez (1948), although extreme, nonetheless exemplifies the
problems inherent in most classifications of the Curimatidae.
Phenetically distinctive species or species groups were
typically recognized as separate genera without consideration
of the consequence of such actions on the phylogenetic
naturalness (sensu Wiley, 1981:71) of previously recognized
taxa. Residual species, under such procedures, in most
instances formed non-monophyletic generic assemblages not
delimited by uniquely derived characters. Rather, those taxa
were characterized by combinations of primitive characters
(symplesiomorphies) and did not delimit complete evolution-
ary lineages. As such they did not constitute valid units for
comparative studies: morphological, biogeographic, or other-
wise.
The studies reported in this paper were undertaken in order
to advance a uniform hypothesis of higher level phylogenetic
relationships within the Curimatidae based on discrete derived
characters. Such a framework will provide a comparative
context for future studies of the family, serving as the basis for
comprehensive revisionary studies of subunits of the family
and analyses of biogeographic patterns within the Curimatidae.
ACKNOWLEDGMENTS.?I am greatly indebted to the follow-
ing individuals and institutions for the loan and exchange of
species, information, and other types of assistance: the late Dr.
Donn E. Rosen, Dr. Gareth J. Nelson, and Dr. C. Lavett Smith,
American Museum of Natural History (AMNH); Dr. P.
Humphry Greenwood and Mr. Gordon J. Howes, British
Museum (Natural History) (BMNH); Dr. William Eschmeyer
and Ms. Pearl Sonoda, California Academy of Sciences (CAS);
Dr. Karel Liem, Dr. William L. Fink, and Mr. Karsten Hartel,
Museum of Comparative Zoology; Dr. Naercio Menezes, Dr.
Heraldo Britski, and Dr. Jose Lima de Figueiredo, Museo de
Zoologia da Universidade de Sao Paulo (MZUSP); Dr. Antonio
Machado-Allison and Dr. Francisco Mago-Leccia, Museo de
Biologia, Universidad Central de Venezuela, Caracas; Dr. Sven
O. Kullander, Swedish Museum of Natural History, Stock-
holm; Prof. Hernan Ortega, Universidad Nacional Mayor de
San Marcos, Lima; Dr. Michael Goulding, Museu Paraense
"Emilio Goeldi," Belem; Dr. Gustavo W. Nunan, Museu
Nacional, Rio de Janiero; and Dr. Michel Jegu, Instituto
Nacional de Pesquisas da Amazonia, Manaus.
Ms. Susan L. Jewett, Mrs. Ann W. Vari, and Mr. Andrew
G. Gerberich, National Museum of Natural History, Smith-
sonian Institution (NMNH), provided technical assistance
during the project. Histological preparations of the bucco-
pharyngeal complexes and ligamentum primordiale cartilage
body were made by Mrs. Helen Wimer (NMNH). The
Portuguese "Resumo" was prepared by Prof. Ricardo M.C.
Castro (MZUSP). Figures 1-3 and 28-31 were prepared by
Mr. Theophilus B. Griswold (NMNH). The paper benefited
from the comments and criticisms of Dr. Stanley H. Weitzman,
Dr. Wayne C. Starnes, and Mrs. Ann W. Vari (NMNH), Prof.
Ricardo M.C. Castro (MZUSP), Mr. Gordon J. Howes
(BMNH), and Dr. Lynne R. Parenti (CAS). I thank all of the
above for their assistance and interest.
Research associated with this study was supported in part
by the Neotropical Lowland Research Program of the
International Environmental Studies Program of the Smith-
sonian Institution.
Systematic Procedures
Phylogenetic relationships within the Curimatidae are
analysed using the methodology of phylogenetic reconstruc-
tion, first formalized by Hennig (1950, 1966). The goal of
phylogenetic systematics (alternatively termed cladistics or
cladism) is the grouping of taxa in series of nested, hierarchical
units that reflect the best estimate of their evolutionary history.
Taxa are grouped on the basis of the possession of shared
derived characters (synapomorphies). Such characters are
considered the only valid basis for cladistic hypotheses of
common ancestry. Hypotheses of relationship advanced on the
basis of the common possession of shared primitive characters
(symplesiomorphies), together with phylogenetic speculations
based on concepts of overall phenetic similarity, or degrees of
difference are either not useful methods for evaluating
alternative phylogenetic hypotheses, or are incongruent with
the aims of this study?the advancement of hypotheses of
phylogenetic relationships within the Curimatidae. In keeping
with the general scientific principle of parsimony, the
hypothesis of the phylogenetic history of a group that
necessitates the fewest ad hoc assumptions about character
NUMBER 471
transformations is preferred. Here, a computer algorithm,
PAUP (Swofford, 1985), is employed to ensure maximum
parsimony.
This does not assume that evolutionary mechanisms are
necessarily parsimonious, but only that parsimony (simplicity)
is the best available inferential principle (Beatty and Fink,
1979; Wiley, 1981). Monophyletic groups are defined as
consisting of a hypothesized common ancestor of the lineage
and all of its descendants. Traditional alternative definitions
of monophyly (see summaries in Wiley, 1981) are of such
generality as to include all possible combinations of subunits
within a lineage, or are dependent on arbitrary taxonomic
ranks, or both, and are therefore often meaningless in
application and information content.
Character polarity (plesiomorphy vs. apomorphy; primitive
vs. derived) is determined by outgroup comparisons, or via
data from ontogenetic transformations, or both. Vari (1983:46-
50) suggested that the Curimatidae is the sister group of the
Prochilodontidae and the lineage formed by those two families
is most closely related to that consisting of the Anostomidae
and Chilodontidae. For purposes of character polarity determi-
nations within the Curimatidae, outgroup comparisons are
centered first on the Prochilodontidae and then on the
Anostomidae and Chilodontidae. Comparisons to other,
distantly related, characiform groups, both Neotropical and Old
World, were undertaken in instances where outgroup compari-
sons to the three noted phylogenetically proximate families
provided equivocal data on character polarity. Such broader
outgroup comparisons were also carried out when the character
in question in each of the proximate outgroup taxa was, in turn,
uniquely derived relative to the generalized condition of
characiforms and thus an inappropriate approximation of the
hypothesized plesiomorphous condition of the character for
characiformes. Conflicts between polarity decisions based on
outgroup comparisons and those arrived at from ontogenetic
data are analyzed within the context of the overall phylogeny.
Discussions of the theory and methods of character polarity
determination can be found in Wiley (1981), Watrous and
Wheeler (1981), and Maddison et al. (1984).
Methods and Materials
Osteological and cartilaginous skeletal systems were exam-
ined in specimens cleared and counterstained for cartilage and
bone with alizarin Red S and alcian blue following the methods
outlined by Taylor and Van Dyke (1985). Previously cleared
specimens stained solely in alizarin Red S, along with dry
skeletal materials, were supplemental sources of osteological
data. Anatomical illustrations were prepared using a Zeiss
microscopic camera lucida.
TERMINOLOGY.?Myological terminology follows that of
Winterbottom (1974). Osteological nomenclature is that of
Weitzman (1962) with the following modifications. Vomer is
substituted for prevomer, and intercalar for opisthotic. The
element traditionally termed the epihyal is instead referred to
as the posterior ceratohyal. The ceratohyal of many previous
authors is more accurately termed the anterior ceratohyal. This
shift in terminology is a consequence of the lack of serial
homology between the so-called epihyal and the epibranchials,
and the resultant misleading inference of homology inherent
in the continued use of epihyal (Nelson, 1969:480-481). The
use of epioccipital rather than epiotic follows Patterson (1975).
Nelson's substitution (1973) of angulo-articular for articular
and retroarticular for angular is more reflective of the
homologies of these elements among teleosts than previous
terminologies, and is utilized in this paper.
Unless otherwise noted, the concepts of the characiform
families used in this paper are those of Greenwood et al. (1966)
with three modifications. The Cynodontidae of those authors
is considered a tribe in the Characidae rather than a distinct
family, in keeping with the results of Howes (1976). The
Ichthyboridae of Greenwood et al. (1966) is placed within the
Distichodontidae following Vari (1979). Contrary to most
previous authors, the Curimatidae of this study does not include
the genera Anodus Spix and Eigenmannina Fowler, which are
considered to be members of the Hemiodontidae after Roberts
(1974). The higher level classification of the Ostariophysi
follows Fink and Fink (1981). The nomenclature for the
African members of the Characidae is that proposed by Paugy
(1984:140-183).
The introductory discussion alluded to the complexity of the
taxonomic schemes that have been applied to supraspecific
subunits of the Curimatidae. The confusion engendered by the
availability of over thirty generic level taxa and eight
suprageneric groupings within the family is compounded by
questions of species recognizability, errors in the morphologi-
cal characterization of generic type species by Eigenmann and
Eigenmann (1889b) and Fernandez-Yepez (1948) (see Vari,
1984a: 13-16, 25), and the presence of a number of un-
described species in the family. The resolution of many of these
problems is the subject of revisionary studies published (Vari,
1982a,b, 1984a,b, 1987; Vari and Gery, 1985; Vari and
Nijssen, 1986) and in progress. The numerous monotypic
genera, many of which will not be recognized when revisionary
studies are completed, makes resolution to the level of all
nominal genera unproductive. The following discussion
therefore deals with what would have constituted multigeneric
assemblages under many previous classificatory schemes,
particularly that of Fernandez-Yepez (1948). The reduced
number of recognized groups either represents taxonomic
changes already proposed (Vari, 1982a, 1984a), or reflects the
likely eventual classificatory schemes that will be proposed
on the basis of research in progress. Table 1 lists the generic
complexes used in this paper and the nominal genera that each
incorporates. The assignment of the nominal species in the
Curimatidae to the recognized genera is detailed in Tables 2
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
TABLE 1.?Genera recognized in this study, their included nominal genera, and the type species of those genera.
Authors and date of description given for genera and type species of each genus.
Recognized genus Included nominal genera and their type species Recognized genus Included nominal genera and their type species
Curimatopsis Curimatopsis Steindachner, 1876; Curimatus (Curimatop-
sis) macrolepis Steindachner, 1876
Curimatichthys Fernandez-Yepez, 1948; Curimatopsis
microlepis Eigenmann and Eigenmann, 1889a
Potamorhina Potamorhina Cope, 1878; Curimatus (Anodus) pristi-
gaster Steindachner, 1876
Casterotomus Eigenmann, 1910; Anodus latior Spix,
1829 (in Spix and Agassiz, 1829)
G aster oslomus.?Fernandez- Yepez, 1948 [incorrect subse-
quent spelling of Gasterotomus]
Suprasinelepichthys FemandezYepez, 1948; Curimatus
laticeps Valenciennes, 1849 (in Cuvierand Valencien-
nes, 1849)
Polamorrhina Braga and Azpelicueta, 1983 [unjustified
emendation of Potamorhina}
Curimata Curimata Bosc, 1817; Salmo edentulous Bloch, 1794
Semitapicis Eigenmann and Eigenmann, 1889b; Charax
planirostris Gray, 1854
Peltapleura Fowler, 1906; Salmo cyprinoides Linnaeus,
1766
Acuticurimata Fowler, 1941; Curimatus macrops Eigen-
mann and Eigenmann, 1889b
Allenina Fernandez-Yepez, 1948; Curimata murieli
Allen, 1942 (in Eigenmann and Allen, 1942)
Lambepiedra Femandez-Yepez, 1948; Lambepiedra
alleni Fernandez-Yepez, 1948
Bilricarinata Fernandez-Yepez, 1948; Curimatus schom-
burgkii Gunther, 1864
Bondichthys Whitley, 1953; Curimatus mivartii Steindach-
ner, 1878
Stupens Whitley, 1954; Curimatus simulatus Eigenmann
and Eigenmann, 1889b
Semitapiscis Braga and Azpelicueta, 1983 [unjustified
emendation of Semitapicis].
Psectrogaster Psectrogaster Eigenmann and Eigenmann, 1889a; Psec-
trogasler rhomboides Eigenmann and Eigenmann,
1889a
Steindachnerina
Pseudocurimata
Curimatella
Cyphocharax
Pseudopsectrogaster Fernandez-Yepez, 1948; Psectrogas-
ter curviventris Eigenmann and Kennedy, 1903
Hamatichthys Fernandez-Yepez, 1948; Anodus ciliatus
Miiller and Troschel, 1845
Semelcarinata Fernandez-Yepez, 1948; Curimatus rutiloi-
r, 1859
Steindachnerina Fowler, 1906; Curimatus trachystetus
Cope, 1878
Rivasella Femandez-Yepez, 1948; Curimatus melaniris
Fowler, 1940
Curimatorbis Fernandez-Yepez, 1948; Curimatus atratoen-
sis Eigenmann, 1912b
Pseudocurimata Fernandez-Ydpez, 1948; Curimatus
lineopunctatus Boulenger, 1911
Cwimatella Eigenmann and Eigenmann, 1889b; Curi-
imatus lepidurus Eigenmann and Eigenmann, 1889a
Apolinarella Femandez-Yepez, 1948; Curimatus meyeri
Steindachner, 1882a
Walbaunina Fernandez-Ydpez, 1948; Curimatus dor-
salis Eigenmann and Eigenmann, 1889b
Lepipinna Femandez-Yepez, 1948; Anodus alburnus
Muller and Troschel, 1845
Cyphocharax Fowler, 1906; Curimatus spilurus Gunther,
1864
Xyrocharax Fowler, 1913b; Curimata stigmaturus
Fowler, 1913b
Hemicurimata Myers, 1929; Curimata esperanzae
Myers, 1929
Curimatoides Fowler, 1940; Curimatoides ucayalensis
Fowler, 1940
Cruxentina Femandez-Yepez, 1948; Curimata hypos-
loma hastata Allen, 1942 (in Eigenmann and Allen,
1942)
and 3. Revisionary studies have been published for two of the
c lades, Curimatopsis (Vari, 1982a,b) and Potamorhina (Vari,
1984a), and the nomenclature and synonymies of those
revisions is followed in this paper. Since such revisions have
not yet been published for the other lineages in the family, the
discussion of the species in those taxa will be based on all the
nominal species assignable to those lineages without prejudg-
ing which species will be recognized as valid in future studies.
Each synapomorphy is assigned a number in order to
facilitate cross reference between the descriptive text, the
listing in "Synapomorphy List and Phylogenetic Reconstruc-
tion," and the cladograms (Figures 44, 45), except that
synapomorphies used exclusively within the genus Curimatop-
sis are assigned letters (A-K), for reasons explained in the
text. Synapomorphy numbering is sequential in the final
synapomorphy listing rather than in the descriptive text in order
to simplify the discussion within "Synapomorphy and Phylo-
genetic Reconstruction." The synapomorphies, their polarities,
and other comments are described in greatest detail within
"Character Description and Analysis."
NUMBER 471
TABLE 2.?Genera recognized in this study and their contained nominal species, subspecies, and varieties.
Species, subspecies, and varieties are listed chronologically within genera and are cited as in their original
descriptions, with authors and dates of description.
Recognized genus Included species Recognized genus Included species
Curimatopsis Cwimala (Curimatopsis) macrolepis Steindachner, 1876
Cur'unatopsis microlepis Eigenmann and Eigenmann,
1889a
Curimatopsis macrocephalus Ahl, 1931
Curimalopsis evelynae Gery, 1964a
Curimatopsis crypticus Van, 1982a
Curimalopsis myersi Vari, 1982b
Potamorhina Anodus latior Spix, 1829 (in Spix and Agassiz, 1829)
Curimatus laticeps Valenciennes, 1849 (in Cuvier and
Valenciennes, 1849)
Curimatus (Anodus) pristigaster Steindachner, 1876
Curimatus altamazonicus Cope, 1878
Semitapiscis squamoralevis Braga and Azpelicueta,
1983
Curimata Salmo immaculatus Linnaeus, 1758
Salmo cyprinoides Linnaeus, 1766
Salmo edentulus Bloch, 1794
Charax planirostris Gray, 1854
Curimatus vittatus Kner, 1859
Curimatus schomburgkii Giinther, 1864
Curimatus asper Giinther, 1868
Curimatus knerii Steindachner, 1876
Curimatus mivartii Steindachner, 1878
Curimatus ocellatus Eigenmann and Eigenmann, 1889b
Curimatus simulatus Eigenmann and Eigenmann, 1889b
Curimatus macrops Eigenmann and Eigenmann, 1889b
Curimata copei Fowler, 1906
Curimatus semilaeniatus Steindachner, 1917
Psectrogaster cisandina Allen, 1942 (in Eigenmann and
Allen, 1942)
Curimata murieli Allen, 1942 (in Eigenmann and Allen,
1942)
Lambepiedra alleni Fernandez-Yepez, 1948
Curimata cerasina Vari, 1984b
Curimata incompta Vari, 1984b
Curimata species A (Vari, in press)
Curimata species B (Vari, in press)
Psectrogaster Anodus ciliatus Miiller and Troschel, 1845
Curimatus rutiloides Kner, 1859
Curimatus essequibensis Giinther, 1864
Psectrogaster rhomboides Eigenmann and Eigenmann,
1889a
Psectrogaster amazonica Eigenmann and Eigenmann,
1889a
Curimatus falcatus Eigenmann and Eigenmann, 1889b
Curimatus isognathus Eigenmann and Eigenmann,
1889b
Pseclrogaster auratus Gill, 1895
Psectrogaster curviventris Eigenmann and Kennedy,
1903
Curimata pearsoni Myers, 1929
Curimata saguiru Fowler, 1941
Psectrogaster rhomboides auslrale Risso and Sanchez,
1964
Sleindachnerina Curimatus argenteus Gill, 1858
Curimatus leuciscus Giinther, 1868
Curimatus dobula Giinther, 1868
Curimatus elegans Steindachner, 1874
Curimatus bimaculatus Steindachner, 1876
Curimatus trachystetus Cope, 1878
Curimatus nasus Steindachner, 1882b
Curimatus hypostoma Boulenger, 1887
Curimatus bimaculatus sialis Eigenmann and Eigen-
mann, 1889b
Curimatus elegans bahiensis Eigenmann and Eigen-
mann, 1889b
Curimatus gilberti brevipinnis Eigenmann and Eigen-
mann, 1889b
Curimatus guntheri Eigenmann and Eigenmann, 1889b
Curimatus conspersus Holmberg, 1891
Curimatus nitens Holmberg, 1891
Curimatus nigrotaenia Boulenger, 1902
Curimatus leuciscus bolivae Eigenmann and Ogle, 1907
Prochilodus stigmaturus Fowler, 1911
Curimatus elegans var. amazonica Steindachner, 1911
Curimatus atratoensis Eigenmann, 1912b
Curimatus issororoensis Eigenmann, 1912a
Curimatus morawhannae Eigenmann, 1912a
Prochilodus pterostigma Fowler, 1913b
Curimatus semiornatus Steindachner, 1914
Curimatus metae Eigenmann, 1922
Curimatus binotatus Pearson, 1924
Curimatus notonotus Miranda Ribeiro, 1937
Curimata melaniris Fowler, 1940
Curimata hypostoma hastata Allen, 1942 (in Eigenmann
and Allen, 1942)
Curimatus robustula Allen, 1942 (in Eigenmann and
Allen, 1942)
Curimata niceforoi Fowler, 1943
Allenina pectinata Femandez-Yepez, 1948
Cruxentina insculpta Femandez-Yepez, 1948
Curimata fasciata Vari and Gery, 1985
Curimata stigmosa Vari, 1987
Curimata biornata, Braga and Azpelicueta, 1987
Pseudocurimata Anodus troschelii Giinther, 1859
Curimatus boulengeri Eigenmann, 1907 (in Eigenmann
and Ogle, 1907)
Curimatus brevipes Eigenmann and Ogle, 1907
Curimatus aureus Pellegrin, 1908
Curimatus lineopunclatus Boulenger, 1911
Curimatus patiae Eigenmann, 1914 (in Eigenmann,
Henn, and Wilson, 1914)
Curimatus peruanus Eigenmann, 1922
Curimatella Anodus alburnus Miiller and Troschel, 1845
Curimatus Meyeri Sleindachner, 1882b
Curimatus dorsalis Eigenmann and Eigenmann, 1889b
Curimatus lepidurus Eigenmann and Eigenmann, 1889a
Curimatus leucoslictus Eigenmann and Eigenmann,
1889b
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
TABLE 2.?Continued.
Recognized genus Included species
Curimatus serpae Eigenmann and Eigenmann, 1889a
Curimatus alburnus australe Eigenmann and Kennedy,
1903
Curimatus elegans paraguayensis Eigenmann and Ken-
nedy, 1903
Curimatus (Curimatella) xinguensis Steindachner, 1908
Curimatus (Curimatella) alburnus var. caudimaculata
Pellegrin, 1909
Curimatus bolivarensis Steindachner, 1910
Curimata reticulata Allen, 1942 (in Eigenmann and
Allen, 1942)
Le pi pinna immaculata Fernandez-Yepez, 1948
Cyphocharax Curimata Gilbert Quoy and Gaimard, 1824
Curimatus abrarno'uks Kner, 1859
Curimatus spilurus Giinther, 1864
Curimatus voga Hensel, 1869
Curimatus albula Liitken, 1874
Curimatus Magdalenae Steindachner, 1879
Curimatus plat anus Giinther, 1880
Curimatus Nagelii Steindachner, 1882a
Curimatus microcephalus Eigenmann and Eigenmann,
1889b
Curimatus plumbeus Eigenmann and Eigenmann, 1889b
Recognized genus Included species
Curimatus spiluropsis Eigenmann and Eigenmann,
1889b
Curimatus gillii Eigenmann and Kennedy, 1903
Curimatus notatus Steindachner, 1908
Curimatus helleri Steindachner, 1910
Curimatus surinamensis Steindachner, 1910
Curimatus stigmaturus Fowler, 1913b
Curimatus multilineatus Myers, 1927
Curimata esperanze Myers, 1929
Curimatus hermanni Ahl, 1931
Curimatella rehni Fowler, 1932
Curimatopsis saladensis Meinken, 1933
Curimatopsis maculatus Ahl, 1934
Curimatoides ucayalensis Fowler, 1940
Curimatus Vandeli Puyo, 1943
Pseudocurimata grandocule Fernandez-Y6pez, 1948
Curimatorbis modestus Fernandez-Y6pez, 1948
Pseudocurimata santacatarinae Femandez-Yepez, 1948
Pseudocurimata steindachneri Femlndez-Yepcz, 1948
Curimatus (Hemicurimata) esperanzae pijpersi Gery,
1965
Curimata vanderi Britski, 1980
Curimata punclata Vari and Nijssen, 1986
Curimata spilota Vari, 1987
TABLE 3.?Nominal species within the Curimatidae, their describers and date of description, together with the
genera to which they are assigned in this study. Species are arranged alphabetically by specific, subspecific, or
varietal epithet, and are cited as in their original description. Undescribed species (e.g., species A, B) are not
included.
Species, author, and date of description
Curimatus abramoides Kner, 1859
Curimatus albula Liitken, 1874
A nodus alburnus Miiller and Troschel, 1845
Lambepiedra alleni Femandez-Yepez, 1948
Curimatus altamazonicus Cope, 1878
Psectrogaster amazonica Eigenmann and Eigenmann,
1889a
Curimatus elegans var. amazonica Steindachner, 1911
Curimatus argenteus Gill, 1858
Curimalus asper Giinther, 1868
Curimatus atratoensis Eigenmann, 1912b
Psectrogaster auratus Gill, 1895
Curimatus aureus Pellegrin, 1908
Curimatella alburnus australe Eigenmann and Ken-
nedy, 1903
Psectrogaster rhomboides australe Risso and Sanchez,
1964
Curimatus elegans bahiensis Eigenmann and Eigen-
mann, 1889b
Curimatus bimaculatus Steindachner, 1876
Curimatus binotatus Pearson, 1924
Curimala biornata Braga and Azpelicueta, 1987
Curimatus leuciscus bolivae Eigenmann and Ogle, 1907
Curimatus bolivarensis Steindachner, 1910
Genus
Cyphocharax
Cyphocharax
Curimatella
Curimata
Potamorhina
Psectrogaster
Steindachnerina
Steindachnerina
Curimala
Steindachnerina
Psectrogaster
Pseudocurimata
Curimatella
Psectrogaster
Steindachnerina
Steindachnerina
Steindachnerina
Steindachnerina
Steindachnerina
Curimatella
Species, author, and date of description
Curimatus boulengeri Eigenmann, 1907 (in Eigenmann
and Ogle, 1907)
Curimatus brevipes Eigenmann and Ogle, 1907
Curimatus gilberti brevipinnis Eigenmann and Eigen-
mann, 1889b
Curimatus alburnus var. caudimaculatus Pellegrin, 1909
Curimata cerasina Vari, 1984b
Anodus ciliatus Miiller and Troschel, 1845
Psectrogaster cisandina Allen, 1942 (in Eigenmann and
Allen, 1942)
Curimatus conspersus Holmberg, 1891
Curimata copei Fowler, 1906
Curimatopsis cryplicus Vari, 1982a
Psectrogaster curviventris Eigenmann and Kennedy,
1903
Salmo cyprinoides Linnaeus, 1766
Curimatus dobula Giinther, 1868
Curimatus dorsalis Eigenmann and Eigenmann, 1889b
Salmo edentulus Bloch, 1794
Curimatus elegans Steindachner, 1874
Curimata esperanzae Myers, 1929
Curimatus essequibensis Gunther, 1864
Curimatopsis evelynae Gery, 1964a
Curimatus falcatus Eigenmann and Eigenmann, 1889b
Genus
Pseudocurimata
Pseudocurimata
Steindachnerina
Curimatella
Curimata
Psectrogaster
Curimata
Steindachnerina
Curimata
Curimatopsis
Psectrogaster
Curimata
Steindachnerina
Curimatella
Curimata
Steindachnerina
Cyphocharax
Psectrogaster
Curimatopsis
Psectrogaster
NUMBER 471
TABLE 3.?Continued.
Species, author, and date of description
Curimata fasciata Van and Gery, 1985
Cwimata Gilbert Quoy and Gaimard, 1824
Curimatus gillii Eigenmann and Kennedy, 1903
Pseudocurimata grandocule Fernandez-Yepez, 1948
Curimatus guntheri Eigenmann and Eigenmann, 1889b
Curimata hypostoma hastala Allen, 1942 (in Eigenmann
and Allen, 1942)
Curimatus helleri Steindachner, 1910
Curimatus hermanni Ahl, 1931
Curimatus hypostoma Boulenger, 1887
Lepipinna immaculata Fernandez-Yepez, 1948
Salmo immaculatus Linnaeus, 1758
Curimata incompta Van, 1984b
Cruxentina insculpta Fernandez-Yepez, 1948
Curimatus isognathus Eigenmann and Eigenmann,
1889b
Curimatus issororoensis Eigenmann, 1912a
Curimatus knerii Steindachner, 1876
Curimatus laticeps Valenciennes, 1849 (in Cuvier and
Valencinnes, 1849)
Anodus latior Spix, 1829 (in Spix and Agassiz, 1829)
Curimatus lepidurus Eigenmann and Eigenmann, 1889a
Curimatus leuciscus Giinther, 1868a
Curimatus leucostictus Eigenmann and Eigenmann,
1889b
Curimatus lineopunctatus Boulenger, 1911
Curimatopsis macrocephalus Ahl, 1931
Curimata (Curimatopsis) macrolepis Steindachner, 1876
Curimatus macrops Eigenmann and Eigenmann, 1889b
Curimatopsis maculatus Ahl, 1934
Curimatus Magdalenae Steindachner, 1879
Curimata melanira Fowler, 1940
Curimatus metae Eigenmann, 1922
Curimatus Meyeri Steindachner, 1882b
Curimatus microcephalus Eigenmann and Eigenmann,
1889b
Curimatopsis microlepis Eigenmann and Eigenmann,
1889a
Curimalus mivartii Steindachner, 1878
Curimatorbis modest us Fernandez-Yepez, 1948
Curimatus morawhannae Eigenmann, 1912a
Curimatus multUineatus Myers, 1927
Curimata murieli Allen, 1942 (in Eigenmann and Allen,
1942)
Curimatopsis myersi Van, 1982b
Curimatus Nagelii Steindachner, 1882a
Curimatus nasus Steindachner, 1882b
Curimata niceforoi Fowler, 1943
Curimatus nigrotaenia Boulenger, 1902
Curimatus nitens Holmberg, 1891
Curimatus notatus Steindachner, 1908
Curimatus notonotus Miranda-Ribiero, 1937
Curimatus ocellatus Eigenmann and Eigenmann, 1889b
Genus
Sleindachnerina
Cyphocharax
Cyphocharax
Cyphocharax
Steindachnerina
Sleindachnerina
Cyphocharax
Cyphocharax
Steindachnerina
Curimatella
Curimata
Curimata
Steindachnerina
Psectrogaster
Steindachnerina
Curimata
Potamorhina
Potamorhina
Curimatella
Sleindachnerina
Curimatella
Pse udocurimata
Curimatopsis
Curimatopsis
Curimata
Cyphocharax
Cyphocharax
Sleindachnerina
Sleindachnerina
Curimatella
Cyphocharax
Curimatopsis
Curimata
Cyphocharax
Steindachnerina
Cyphocharax
Curimata
Curimatopsis
Cyphocharax
Sleindachnerina
Steindachnerina
Sleindachnerina
Steindachnerina
Cyphocharax
Steindachnerina
Curimata
Species, author, and date of description
Curimatus elegans paraguayensis Eigenmann and Ken-
nedy, 1903
Curimatus patiae Eigenmann, 1914 (in Eigenmann,
Henn, and Wilson, 1914)
Curimata pearsoni Myers, 1929
Allenina pectinata Fernandez-Yepez, 1948
Curimatus peruanus Eigenmann, 1922
Curimatus (Hemicurimata) esperanzae pijpersi Gery,
1965
Charax planirostris Gray, 1854
Curimatus platanus Giinther, 1880
Curimatusplumbeus Eigenmann and Eigenmann, 1889b
Curimatus (Anodus) pristigaster Steindachner, 1876
Prochilodus pterostigma Fowler, 1913b
Curimata punctata Van and Nijssen, 1986
Curimatella rehni Fowler, 1932
Curimata reticulata Allen, 1942 (in Eigenmann and
Allen, 1942)
Psectrogaster rhomboides Eigenmann and Eigenmann,
1889a
Curimata robustula Alien, 1942 (in Eigenmann and
Allen, 1942)
Curimatus rutiloides Kner, 1859
Curimata saguiru Fowler, 1941
Curimatopsis saladensis Meinken, 1933
Pseudocurimata santacatarinae Fernandez-Yepez, 1948
Curimatus schomburgkii Giinther, 1864
Curimatus semiornatus Steindachner, 1914
Curimatus semilaeniatus Steindachner, 1917
Curimatus serpae Eigenmann and Eigenmann, 1889a
Curimatus bimaculatus sialis Eigenmann and Eigen-
mann, 1889b
Curimatus simulalus Eigenmann and Eigenmann, 1889b
Curimata spilota Vari, 1987
Curimatus spiluropsis Eigenmann and Eigenmann,
1889b
Curimatus spilurus Giinther, 1864
Semitapiscis squamoralevis Braga and Azpelicueta,
1983
Pseudocurimata steindachneri Fernandez-Yepez, 1948
Curimata stigmosa Vari, 1987
Prochilodus stigmaturus Fowler, 1911
Curimatus stigmaturus Fowler, 1913a
Curimatus surinamensis Steindachner, 1910
Curimatus trachystetus Cope 1878
Anodus troschelii Giinther, 1859
Curimatoides ucayalensis Fowler, 1940
Curimatus Vandeli Puyo, 1943
Curimatus vanderi Britski, 1980
Curimatus vittatus Kner, 1859
Curimatus voga Hensel, 1869
Curimatus (Curimatella) xinguensis Steindachner, 1908
Genus
Curimatella
Pseudocurimata
Psectrogaster
Steindachnerina
Pseudocurimata
Cyphocharax
Curimata
Cyphocharax
Cyphocharax
Potamorhina
Steindachnerina
Cyphocharax
Cyphocharax
Curimatella
Psectrogaster
Steindachnerina
Psectrogaster
Psectrogaster
Cyphocharax
Cyphocharax
Curimata
Steindachnerina
Curimata
Curimatella
Steindachnerina
Curimata
Cyphocharax
Cyphocharax
Cyphocharax
Potamorhina
Cyphocharax
Steindachnerina
Sleindachnerina
Cyphocharax
Cyphocharax
Steindachnerina
Pseudocurimata
Cyphocharax
Cyphocharax
Cyphocharax
Curimata
Cyphocharax
Curimatella
NOTE: Curimatus alberti cited by Giinther (1880) and Eigenmann and
Eigenmann (1889b) is presumably a lapsus for Cypocharax gilberti (Quoy and
Gaimard, 1824). Curimatus hasemanni referred to by Fernandez-Yepez
(1948:73) is evidently an incorrect citation of Curimatus hermanni Ahl (1931).
As noted by Gery (1972:33), Curimatus tigris Fowler (1913a) is actually a
member of the Prochilodontidae. Vari and Castro (1988) demonstrate that
Prochilodus stigmalurus Fowler (1911) is a species of Steindachnerina.
10 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
MATERIAL EXAMINED.?Osteological and myological mate-
rials examined are deposited in the following museums: Museo
de Biologia, Universidad Central de Venezuela, Caracas
(MBUCV); the Museum of Comparative Zoology, Harvard
University, Cambridge (MCZ); the Museo de Zoologia da
Universidade de Sao Paulo (MZUSP); and the National
Museum of Natural History, Smithsonian Institution, Washing-
ton, D.C. (USNM).
The following specimens are the basis for text illustrations
or specific observations noted in the text. Measurements, in
mm, are of standard length (SL).
ANOSTOMIDAE
Abramites hypselonotus (Gunther), USNM 164036; 1 specimen, 100.5 mm;
Ecuador, lower Rio Bobonaza.
Anostomus plicatus Eigenmann, USNM 225396; 2 specimens, 75.8-95.3
mm; Surinam, Nickerie District, Matappi Creek.
Anostomus species, USNM 231540; 1 specimen, 81.3 mm; no locality data.
Gnathodolus bidens Myers, USNM 231539; 1 specimen, 76.2 mm; aquarium
material.
Laemolyta species, USNM 179514; 2 specimens, 75.3-77.0 mm; Brazil,
Amazonas, Rio Urubu.
Leporellus vittatus Valenciennes, USNM uncatalogued, 1 specimen, 121.4
mm; Venezuela, Barinas, Rio Las Palmas.
Rhytiodus microlepis Kner, USNM 163850; 1 specimen, 131.7 mm; Peru,
Loreto, Iquitos.
Schizodon fasciatum Agassiz, USNM 179507; 1 specimen, 49.3 mm; Brazil,
Amazonas, Rio Urubu.
Synaptolaemus cingulatus Myers and Femandez-Yepez, MBUCV V-4252; 1
specimen, 71.2 mm; Venezuela, Rio Paragua.
CHARACIDAE
Agoniates species, USNM 243222; 1 specimen, 65.6 mm; Brazil, Amazonas,
mouth of Rio 19a.
Alestes laleralis Boulenger, USNM 285664; 1 specimen, 75.0 mm; Botswana,
Xuguna.
Alestes longipinnis (Gunther), USNM 285665; 3 specimens, 44.7-50.5 mm;
Togo, Togble-Kope.
Brycon falcatus Muller and Troschel, USNM 226161; 2 specimens, 71.3-78.3
mm; Surinam, Nickerie District, Corantijn River (Figures 6, 18).
Calaprion rnento Muller and Troschel, USNM 257547; 1 specimen, 55.2 mm;
Venezuela, Estado Apure, Rio Cunaviche.
Clupeacharax anchovoides Pearson, USNM 243223; 2 specimens; 57.6-58.7
mm; Peru, Ucayali, Pucallpa.
Exodon paradoxus Muller and Troschel, USNM uncatalogued; 1 specimen,
69.3 mm; Brazil, Goiis, Rio Tocantins.
Hollandichtkys sp., USNM 285668; 3 specimens, 53.5-71.8 mm; no locality
data.
Uydrocynus species, USNM 285669; 3 specimens, 38.5-59.2 mm; Volta
[Ghana], Black Volta River.
Lepidarchus adonis Roberts, USNM 267290; 5 specimens, 16.8-21.3 mm;
Ghana, Aluku.
Phenacogrammus pabrensis Roman, USNM 285666; 2 specimens, 31.0-33.7
mm; Ghana, Dayi River at Vakpo.
CHILODONTIDAE
Caenoiropus labyrinthicus (Kner), USNM 231544; 1 specimen. 64.2 mm;
Venezuela, Territorio Federal Amazonas, upper Rio Orinoco at Tama Tama.
Caenotropus labyrinthicus (Kner), USNM 231543; 1 specimen, 58.5 mm;
Brazil, Amazonas, Rio Negro.
Caenotropus maculosus (Eigenmann), USNM 231545; 2 specimens, 42.7-46.3
mm; Guyana.
Chilodus punctatus Muller and Troschel, USNM 231542, 13 specimens,
27.1 -38.2 mm; Peru, Loreto, Rio Nanay.
CTTHARINIDAE
Citharinus citharus Geoffroy, USNM 52146; 1 specimen, 218.7 mm; Egypt,
Nile River.
CURIMAT1DAE
Cyphocharax abramoides (Kner), USNM 267952, 1 specimen, 110.3 mm;
Brazil, Para, Rio Xingu, Belo Monte (Figures 2, 32).
Cyphocharax helleri (Steindachner), USNM 220158; 3 specimens, 48.7-66.3
mm; Surinam, Brokopondo District, Gran Kreek, 63 km from Afobaka.
Cyphocharax magdalenae (Steindachner), USNM 220197; 2 specimens,
78.7-112.3; Panama, Chiriqui, creek 3 mi [4.8 km] W of San Juan on
Inter-American Highway.
Cyphocharax microcephala (Eigenmann and Eigenmann), USNM 225307; 2
specimens, 28.7-49.3 mm; Surinam, Nickerie Distict, Kapoeri Creek.
Curimata cisandina (Allen), USNM 229171, 1 specimen, 89.9 mm; Brazil,
Amazonas, Rio Solimoes, Parana da Ilha Marchantaria (Figure 19).
Curimata cyprinoides (Linnaeus), USNM 231433; 2 specimens, 17.1-21.4
mm; Surinam, Nickerie District, Corantijn River (Figure 7).
Curimata cyprinoides (Linnaeus), USNM 267964; 1 specimen, 115 mm;
Brazil, Amapa, Rio Araguari, Ferreira Gomes (Figures 8, 22, 23, 26c, 27c,
36c, 42D).
Curimata cyprinoides (Linnaeus), USNM 225619; 1 specimen, 79.2 mm;
Surinam, Nickerie District, Makilikabroe Creek (Figure 14).
Curimata cyprinoides (Linnaeus), USNM 267963; 2 specimens, 110.5 -125.0
mm; Brazil, Para, Rio Xingu, Belo Monte (Figures 28, 30).
Curimata incompta Vari, USNM 273308; 1 specimen, 81.2 mm; Venezuela,
Estado Bolivar, small isolated carlo normally draining into Rio Orinoco,
downstream of El Burro (Figure 3).
Curimata ocellata Eigenmann and Eigenmann, USNM 267973; 1 specimen,
160.3 mm; Brazil, Amazonas, Rio Marauia (Figure 1).
Curimatella alburna (Muller and Troschel), MZUSP 6309; 1 specimen, 98.3
mm; Brazil, Amazonas, mouth of Rio Purus, Lago Castro (Figures 13, 39).
Curimatella meyeri (Steindachner), USNM 261508; 1 specimen, 100.1 mm;
Peru, Ucayali, Rio Ucayali, Masisea (Figure 35c).
Curimatopsis microlepis Eigenmann and Eigenmann, MZUSP 21053; 1
specimen, 96.0 mm; Brazil, Amazonas, Rio Negro (Figures 5, 15).
Curimatopsis microlepis Eigenmann and Eigenmann, USNM 268867; 1
specimen, 75.8 mm; Brazil, Amazonas, Rio Solimoes near Benin (Figures
12, 22A, 26, 27B , 38A, 39, 40A, 42).
Polamorhina altamazonica (Cope), USNM 257367; 1 specimen, 119.8 mm;
Peru, Loreto, Rio Amazonas, near Iquitos (Figures 33A, 3 5 B , 36B, 42B).
Polamorhina laticeps (Valenciennes), USNM 121325; 1 specimen, 129.3 mm;
Venezuela, Zulia, Lago Maracaibo basin (Figures 10, 2 2 B , 41A) .
Potamorhina squamoralevis (Braga and Azpelicueta), USNM 243228; 1
specimen, 90.2 mm; Brazil, Mato Grosso, Baia do Buritizal (Figures 21,25).
Psectrogaster amazonica Eigenmann and Eigenmann, USNM 261518, 1
specimen, 106.0 mm; Peru, Ucayali, Rio Ucayali, PucaUpa (Figures 22c,
36A).
Psectrogaster ciliata (Muiiller and Troschel), USNM 269990; 1 specimen,
94.8 mm; Venezuela, Estado Bolivar, small isolated cano normally draining
into Rio Orinoco, downstream of El Burro (Figures 2 6 B , 38B , 41B, 42C).
Psectrogaster curviventris Eigenmann and Kennedy, USNM 243221; 1
specimen, 72.9 mm; Brazil, Mato Grosso, Baia do Buritizal (Figures 9, 35A).
Pseudocurimata boulengeri (Eigenmann), USNM 285671; 4 specimens,
83.4-92.7 mm; Ecuador, Vinces.
Pseudocurimata lineopunclata (Boulenger), MCZ 54029; 1 specimen, 78.1
NUMBER 471 11
mm; Ecuador, Esmeraldas, Estero La Boveda, 4 km from Camerones
(Athahualpa).
Pseudocurimala patiae (Eigenmann), USNM 285672; 1 specimen, 82.1 mm;
Colombia, Rio Patia, Barbacoas.
Pseudocurimata peruanus (Eigenmann), USNM 285667; 1 specimen, 90.5
mm; Peru, Piura, Tinajones (Figure 43c).
Pseudocurimata troscheli Gunther, USNM 285673; 1 specimen, 80.5 mm;
Ecuador, Colimes.
Steindachnerina argentea (Gill), USNM 285663; 1 specimen, 83.5 mm;
Trinidad, northern Trinidad, Arouca River, just north of Churchill-Roosevelt
Highway (Figure 40B) .
Steindachnerina bimaculata (Steindachner), USNM 251450; 2 specimens,
61.5-75.4 mm; Peru, Loreto, Rio Amazonas, opposite Tabatinga, Brazil
(Figures 33B, 34B, 37).
Steindachnerina conspersa (Holmberg), USNM 232224; 1 specimen, 83.2
mm; Paraguay, Presidente Hayes, off Trans-Chaco Highway at km 50
(Figure 36D).
Steindachnerina hypostoma (Boulenger), USNM 167802; 1 specimen, 82.7
mm; Peru, Rio Huallaga (Figures 11, 20, 24).
Steindachnerina hypostoma (Boulenger), USNM 261493; 1 specimen, 76.8
mm; Peru, Ucayali, Pucallpa (Figure 29).
Steindachnerina hypostoma (Boulenger), USNM 261513; 1 specimen, 73.4
mm; Peru, Ucayali, Rio Ucayali, Pucallpa (Figure 31).
DlSTICHODON TIDAE
Mesoborus species, USNM 285674, 1 specimen, 74.5 mm; no locality data.
Nannocharaxintermedius Boulenger,USNM231555; 2 specimens,50.7-63.4
mm; West Africa.
Paradistichodus dimidiatus Pellegrin, USNM 231556; 2 specimens, 45.6-47.3
mm; Ghana, Dayi River.
Xenocharax spilurus Gunther, USNM 227693; 1 specimen, 89.3 mm; Gabon,
Lac Ezanga (Figure 4).
HEMIODONTIDAE
Anodus elongatus Spix, USNM 231550; 1 specimen, 120.3 mm; Peru, Loreto,
Rio Ucayali.
Argonectes scapularis Bohlke and Myers, USNM 243224; 1 specimen, 125.4
mm; Brazil, Amazonas, Rio Janauperi.
Bivibranchia bimaculata Vari, USNM 225974; 1 specimen, 97.0 mm; Surinam,
Nickerie District, Corantijn River.
Bivibranchia protractila Eigenmann, USNM 194363; 1 specimen, 62.8 mm;
Brazil, Mato Grosso, upper Rio Juruena.
Hemiodopsis ocellala Vari, USNM 225593; 1 specimen, 99.6 mm; Surinam,
Nickerie District, Corantijn River.
Hemiodus species, USNM 231551; 2 specimens, 55.7-57.1 mm; Brazil, Mato
Grosso, Rio Arinos.
Micromischodus sugillatus Roberts, USNM 205527; 1 specimen, 96.1 mm;
Brazil, Para.
HEPSETIDAE
Hepsetus odoe (Bloch), USNM 231553; 1 specimen, 96.2 mm; Togo, Kama.
PARODON TIDAE
Parodon suborbilalis Valenciennes, USNM 231552; 2 specimens, 55.0-58.1
mm; Colombia, Rio Salado.
Saccodon dariensis (Meek and Hildebrand), USNM 208505; 1 specimen, 73.3
mm; Panama, Rio Membrillo.
PROCHILODONTIDAE
Ichthyoelephas species, USNM 231437; 1 specimen, 110.2 mm; Ecuador.
Prochilodus nigricans Agassiz, USNM 231438; 1 specimen, 144.3 mm;
Bolivia, Tumpasa (Figures 6, 16).
Prochilodus rubrotaeniatus Schomburgk, USNM 225419; 1 specimen, 108.7
mm; Surinam, Nickerie District, Corantijn River (Figure 27A).
Semaprochilodus laticeps Steindachner, USNM 270239; 1 specimen, 68.5
mm SL; Venezuela, Estado Bolivar, carlo off Rio Orinoco at El Burro (Figure
34A).
In addition to the listed specimens, a large number of dry
skeletons, stained and cleared glycerine preparations, and
alcohol preserved specimens of nearly all species of the family
Curimatidae were examined in the comparative studies
associated with the present analysis. The majority of those
collections are deposited in the collections of the AMNH,
MBUCV, MZUSP, and USNM, with lesser numbers of
specimens being from the holdings of the other institutions
noted under "Acknowledgments."
ABBREVIATIONS.?The following abbreviations are used in
the text and figures.
AA
AC
AD
AE
AI
ALP
AS
ASE3
ASMQF
AVPE4
BB
BC
BH
BHC
BHTP
BR
C
CART
CB-LP
CCH3
DCE
DEN
DFPB3
DH
DLF
E
ECT
EPC
FP
GA
H
HC-IO4
HY
IL
IO
LF
LIG
LP
angulo-articular
anterior ceratohyal
anterodorsal laterosensory canal segment of sixth infraorbital
anterior extension of ventral process of third hypobranchial (Hj)
area of insertion of ligamentum primordiale
anterior lobulate processes of buccopharyngeal complex
anterior spur of fourth epibranchial (E4)
anterior articular surface of third epibranchial (E3)
anterior section of metapterygoid-quadrate fenestra
anteroventral portion of fourth epibranchial (E4)
basibranchial (1 to 4)
buccopharyngeal complex
basihyal
basihyal cartilage
basihyal tooth plate
branchiostegal rays (1 to 4)
ceratobranchial (1 to 5)
cartilage
cartilage body of ligamentum primordiale
common posterior cartilage of third hypobranchials (H3)
distal cartilage of epibranchial (3 or 4)
dentary
dorsal flange of third infrapharyngobranchial (PB3)
dorsal hypophyal
dorsolateral flange of second infrapharyngobranchial (PB2)
epibranchial (1 to 5)
ectopterygoid
ethmopalatine cartilage
first proximal radial pterygiophore of dorsal fin
gill arch
hypobranchial (1 to 3)
horizontal laterosensory canal of fourth infraorbilal (IO4)
hyomandibula
insertion point of ligament between second and third hypobran-
chials (H2 and H3)
infraorbilal (1 to 6)
lateral fold of buccopharyngeal complex
ligament
ligamentum primordiale
12 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
LPE3
LR
MBS
MES
MET
METR
MF
MPE3
MPL
MQF
MR
MVL
MX
N
NS
PAL
PB
PC
PD
PDAS
PE
PLP
PMX
POP
PP
PSMQF
QU
RA
SF
SN
SPH
SUOP
SYM
TCBB3
UL
UN
UP
VAP
VH
VLF
VPE4
VPH,
lateral articular process of third epibranchial (E3)
lateral ridge on dorsal surface of fourth ceratobranchial (C^)
medial bony spur of fourth epibranchial (E4)
mesopterygoid
metapterygoid
metapterygoid ridge
median fleshy fold of buccopharyngeal complex
medial articular process of third epibranchial (E3)
mucous producing layer of buccopharyngeal complex
metapterygoid-quadrate fenestra
medial ridge on dorsal surface of fourth ceratobranchial (C4)
mid-ventral ligament
maxilla
notch of premaxilla receiving rear portion of premaxilla
neural spines (1 to 7)
palatine
infrapharyngobranchial (1 to 4)
posterior ceratohyal
posterodorsal laterosensory canal segment of sixth infraorbital
posterodorsal articular surface of palatine
posterior extension of ventral process of third hypobranchial (H3)
posterior lobulate process of buccopharyngeal complex
premaxilla
preopercle
posterior process of maxilla
posterior section of metapterygoid-quadrate fenestra
quadrate
retroarticular
secondary fleshy folds of buccopharyngeal complex
supraneurals (1 to 6)
sphenotic
supraopercle
symplectic
terminal cartilage of third basibranchial (BB3)
upper lip
uncinate process
upper pharyngeal tooth plate (4 or 5)
ventral articular process of fourth epibranchial (E4)
ventral hypohyal
ventrolateral flange of second infrapharyngobranchial (PBj)
ventral plate of fourth epibranchial (E^
ventral process of third hypobranchial (H3)
Character Description and Analysis
The primary focus of the present study is the advancement
of an explicit hypothesis of the phylogenetic relationships
within the Curimatidae. Characters that were discovered to
demonstrate phylogenctically significant variation are dis-
cussed in this section, along with their taxonomic distributions
and the information from outgroup and ontogenctic studies
that permits their polarization.
The evidence from the different analyzed characters is
brought together under "Synapomorphy List and Phylogenetic
Reconstruction" to advance the most parsimonious hypothesis
of phylogenetic relationships within the Curimatidae. That
hypothesis will, in turn, be used as the basis to evaluate the
utility of previous intrafamilial classifications as indicators of
the phylogenetic history of the subunits of the family. Since
the aim of the present study is the derivation of an hypothesis
of the evolutionary history within the Curimatidae, I will not
present a detailed description of the osteology and soft anatomy
of the involved lineages. Only anatomical systems that show
variation relevant to the intrafamilial phylogenetic reconstruc-
tion are discussed.
In a few instances, characters pertinent to the question of the
monophyly of the lineage formed by the Curimatidae and
Prochilodontidae, or relevant to hypotheses proposed in
previous studies (Vari, 1982a, 1983,1984a) are also discussed.
Character variation at levels less inclusive than the genus is
not usually detailed. Exceptions involve the genera Curimatop-
sis and Potamorhina which have already been revised (Vari,
1982a and 1984a, respectively) and which were the subjects
of associated phylogenetic analyses. Species level relationships
within the remaining groups will be analyzed as part of
forthcoming revisionary studies.
Among the hypothesized derived characters discovered
during this study, some were found to have distributions within
the Curimatidae incongruent with the most parsimonious
hypothesis of relationships based on the overall distribution
of all synapomorphies. Such homoplasies are not recognizable
a priori, but can only be identified within the context of the
final phylogenetic hypothesis. Homoplasious characters and
the degree to which different groups of characters contribute
to the final sets of congruent synapomorphies and homoplasies
is analyzed in the section "Convergent Characters." Nonethe-
less, in order to simplify the discussion, homoplasies and their
distributional incongruities relative to the hypothesized phylo-
geny will also be discussed at appropriate points in this section.
BRANCHIAL ARCHES
The bones, cartilages, and associated connective tissues and
muscles of the branchial arches have been a source of
considerable morphological information pertinent to the
elucidation of the phylogenetic relationships of the Curimat-
idae. The eight derived branchial basket characters listed by
Vari (1983:11-24, 47) for the Curimatidae and Prochilodon-
tidae represent the majority of the hypothesized synapomor-
phies for that clade. Similarly, half of the characters uniting
the lineage consisting of the Curimatidae and Prochilodontidae
to the clade formed by the Anostomidae and Chilodontidae are
gill arch synapomorphies (Vari, 1983:46-47). Research
associated with this study has uncovered an additional derived
modification in the dorsal portions of the branchial basket
whose phylctic distribution is congruent with a hypothesis of
a sister group relationship between the Curimatidae and
Prochilodontidae along with three additional synapomorphies
for the species of the Curimatidae. This system also
NUMBER 471 13
demonstrates numerous characters that are informative relative
to a hypothesis of generic level relationships within the
Curimatidae.
Dorsal Portion of Branchial Arches
The large number of synapomorphies in the bones,
cartilages, and ligaments of the dorsal portions of the branchial
system within the Curimatidae are discussed below. In order
to simplify discussion of the numerous characters found in the
multitude of individual elements in that system, the analysis
is arranged around subunits of the arches. These subunits
typically consist of individual components (e.g., the fourth
epibranchial and its associated cartilages and ligaments),
although in some situations, functional complexes (e.g., the
dorsal portion of a single branchial arch) will be the unit of
analysis.
The branchial arches of prochilodontids, the hypothesized
sister group to the Curimatidae (Vari, 1983), serve as the best
available approximation of the plesiomorphous condition of
most components of the branchial system for curimatids.
Prochilodontids demonstrate derived gill arch modifications
hypothesized as common to the presumed ancestor of that
family and the Curimatidae. Apart from those few derived
modifications unique to the family, the prochilodontid
branchial basket retains overall its hypothesized ancestral
condition. The Prochilodontidae demonstrates only one modifi-
cation of the condition of the upper portion of the gill arches
hypothesized as common to the ancestors of prochilodontids
and curimatids that is significant for the following discussion.
That character involves the form of the fifth upper pharyngeal
tooth plate (UP5). All members of the Prochilodontidae have
that element restructured into a vertically expanded, trans-
versely flattened plate, an alteration unique for the members
of that family among Characiformes (Vari, 1983:19,49; Figure
6).
As a consequence of its apomorphous nature, the UP5 of the
Prochilodontidae cannot serve as an approximation of the
plesiomorphous condition of that element for comparative
studies within the Curimatidae. Rather it is necessary to use
more distant outgroups in polarizing character state transition
series for that ossification. The sister group to the Prochilodon-
tidae and Curimatidae, the clade formed by the Anostomidae
and Chilodontidae, also has the fifth upper pharyngeal tooth
plate significantly altered autapomorphously (Vari, 1983:19),
making the condition of UP5 in that lineage similarly
inappropriate as an approximation of the plesiomorphous state
of the bone in curimatids. Rather it is necessary to refer to
phyletically more distant characiform taxa as outgroups. The
absence of a definitive hypothesis of a sister group to the
lineage formed by the four families (Curimatidae, Prochilodon-
tidae, Anostomidae, and Chilodontidae) obscures the exact
outgroup most appropriate for outgroup comparisons to that
clade. This ambiguity necessitates reference to all other
characiforms as a general outgroup for comparisons involving
UP5.
Fourth and Fifth Epibranchials (E4, E5)
Vari (1983:48) listed ten synapomorphies for the members
of the Curimatidae (see also "Synapomorphy List and
Phylogenetic Reconstruction"). An eleventh shared derived
character involving the ventral margin of the fourth epibran-
chial of curimatids has been found in the morphological studies
aimed at deriving a hypothesis of phylogenetic relationships
within the family, herein. The anteroventral portion of the
fourth epibranchial (AVPE4, Figure 4) in characiform out-
groups extends more ventrally than the remainder of the bone
and forms a discrete anteroventral process of the element. That
enlarged anterior portion of the bone has two major articular
facets. The broad anterior articular surface contacts the
posterior portion of the fourth infrapharyngobranchial (PB4),
and the smaller ventral articular surface serves as the region
of attachment for the fifth upper pharyngeal tooth plate (UP5)
(e.g., Xenocharax spilurus, Figure 4). The main body of E4
lies posterior and dorsal of this anteroventral articular portion
of the bone, with the two sections of E4 relatively discrete. The
main body of the fourth epibranchial has a transversely rounded
ventral margin that is somewhat thickened relative to the
plate-like triangular portion of the bone that extends dorsal of
it. In curimatids the anteroventral articular portion of the fourth
epibranchial and this posterior horizontal ventrally rounded
section of the bone are not as discrete as in the outgroups. This
fusion, an apomorphous conditon, is the consequence of the
presence of an additional transversely flattened vertical plate
of bone along the ventral portion of the main body of E4
(SYNAPOMORPHY 26). That plate (VPE^ extends along the
ventral margin of the main body of E4 from the posterior
surface of the anteroventral articular process of the bone
(AVPE4) (Figure 5). This plate is horizontally elongate overall
with a somewhat triangular shape. It is most developed
anteriorly where it is continuous along a vertical line with the
posterior surface of the articular process of UP5. The height
of the bony plate gradually tapers posteriorly, with the exact
form and orientation of the plate in different curimatids a
function of the overall morphology of the fourth epibranchial
in those lineages.
The main body of the fourth epibranchial also demonstrates
a notable degree of phylogenetically informative variability
within the Curimatidae. The generalized plan of the bone
among prochilodontids serves well as an example of the
outgroup condition for the Curimatidae. In prochilodontids the
medial surface of the main body of the fourth epibranchial
in the region dorsal of the area where E4 articulates with
14 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
FIGURE 4.?Xenocharax spilurus, USNM 227693, dorsal portion of gill arches, right side, medial view, anterior
to left (dense patterned stippling represents cartilage).
the fifth upper pharyngeal toothplate (UP5) is flat and
unelaborated (Figure 6). The primary axis of the dorsal process
of the fourth epibranchial (the dorsal extension of Nelson,
1967, and the suprapharyngobranchial process of Bertmar et
al., 1969) in prochilodontids has an anterodorsal orientation.
The dorsal process of E4 in that family terminates dorsally in
a transversely flattened cartilage (DCE4) that is relatively
smaller than that in the Curimatidae and widely separated from
the uncinate processes of the more anterior epibranchials (e.g.,
DCE3, Figure 6). A cord-like ligament (LIG) extends between
the posterior surface of the uncinate process of the third
cpibranchial (E3) and the main body of the fourth epibranchial
. That inter-epibranchial connective tissue band inserts in
prochilodontids onto E4 slightly posterolatcral to the area of
articulation of that element with the cartilaginous fourth
infrapharyngobranchial (PB4, Figure 6).
Analysis of the conditions of the fourth epibranchial in
outgroups to the Curimatidae indicates that the form of the
fourth epibranchial described in the preceeding paragraph
represents the plesiomorphous state of the ossification. Those
cited conditions are hypothesized to have been, in turn, the
basis for a number of derived modifications of that ossification
in a number of subunits of the Curimatidae.
The medial surfaces of the main body and dorsal process of
the fourth epibranchial are unelaborated in all examined
characiform outgroups. That simple condition, although also
DCE.
DCE3
MPE3V
DFPBa U P 4 A V P E 4
UR5 VPE4
FIGURE 5.?Curimatopsis microlepis, MZUSP 21053, dorsal portion of gill arches, right side, medial view,
anterior to left (dense patterned stippling represents cartilage).
NUMBER 471 15
DCE,
DFPB3
FIGURE 6.?Prochilodus nigricans, USNM 231438, dorsal portion of gill arches, right side, medial view, anterior
to left (dense patterned stippling represents cartilage).
common to all members of the Prochilodontidae (Figure 6), is
limited within the Curimatidae to the five species of the genus
Curimatopsis (sensu Vari, 1982a,b) (Figure 5). In the genera
Potamorhina, Curimata, Psectrogaster, Steindachnerina, Pseu-
docurimata, Curimatella, and Cyphocharax (Figures 7-11) a
distinct posteroventrally sloping, often distally-tapering bony
spur (MBS), or a further derived form of such a process, arises
from the medial surface of the main body of the fourth
epibranchial. The spur extends medially from the body of E4
to varying degrees towards the midsagital plane, nearly
contacting its counterpart of the other side in some subgroups
of the family. The unique possession of this medial process of
E4 in Potamorhina, Curimata, Psectrogaster, Steindachnerina,
Pseudocurimata, Curimatella, and Cyphocharax among the
diverse characiform taxa examined is evidence of the derived
nature of the possession of the spur. That process is
consequently considered a synapomorphy for these seven
genera (SYNAPOMORPHY 51). Although prochilodontids have
an expansion of the ventral portion of the fourth epibranchial
proximate to the cartilaginous articulation of E4 and UP5 (VAP,
Figure 6), that elaboration differs both in position and form
from the spur on the medial surface of the fourth epibranchial
in Potamorhina, Curimata, Psectrogaster, Steindachnerina,
Pseudocurimata, Curimatella, and Cyphocharax. The structure
in the Prochilodontidae is consequently considered non-
homologous with the process in the cited curimatid genera.
The function of the medial spur arising from the body of the
fourth epibranchial of most curimatids is uncertain. Although
some connective tissue sheets attach to the spur, neither
discrete well-developed ligaments nor muscles arise from that
process. The dorsal surface of the spur does, however, contact
and conform in shape to the proximate ventral region of the
medial and posterior portions of the ventral surface of the
muscular epibranchial organ common to all curimatids.
Therefore the spur may serve, at least partially, as a support
for the large epibranchial organ typical of the family. That
functional hypothesis must be qualified by the observation
that the Prochilodontidae, the sister group to the Curimatidae,
along with the five species of the genus Curimatopsis have
well developed epibranchial organs in the absence of a medial
spur on the fourth epibranchial. Alternatively, the spur may
function to some extent to stiffen a portion of the roof of the
pharyngeal cavity. In the intact branchial basket of most
curimatids the distal portion of the spur is an integral part of
the roof of the pharyngeal cavity and stiffens that muscular
tube directly {Potamorhina, Psectrogaster, Steindachnerina,
Pseudocurimata, Curimatella, and Cyphocharax). In species
of Curimata the spur is not directly integrated into the
pharyngeal roof but rather supports the fifth upper pharyngeal
tooth plate (sec discussion below) which is, in turn,
incorporated into the fleshy pharyngeal roof. Thus in both
conditions the spur functions in reducing the flexibility of the
fleshy upper portion of the pharyngeal cavity, a modification
that presumably assists in maintaining contact between that
region of the pharynx and the opposing dorsal surface of the
fifth ceratobranchial during food item manipulation.
16 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
This medial bony process of the fourth epibranchial
undergoes further apomorphous restructuring in some sub-
groups of non-Curimatopsis curimatids. These modifications
serve to delimit several smaller assemblages within the family.
The twelve nominal species assigned to Psectrogaster have the
overall form of the medial spur of the fourth epibranchial
significantly lengthened posteriorly (Figure 9) relative to the
condition in the other members of the family with the structure.
The unique nature of the elongation of the spur relative to the
medially unelaborated E4 typical of most characiforms, and
with respect to the more moderately developed spur in other
non-Curimatopsis curimatids is congruent with the hypothesis
of the elongation being a synapomorphy for the species of
Psectrogaster (SYNAPOMORPHY 86).
The spur on the fourth epibranchial in the species of
Curimata differs from the form of the process in other
curimatids in undergoing a progressive ontogenetic expansion
ventrally. In the majority of curimatids with a medial spur on
E4 that process has a relatively small base where it arises from
the medial surface of the fourth epibranchial. The ventral
terminus of the base of the spur, in turn, is distinctly separated
from the ventral articular surface of E4. Similarly the ventral
margin of the distal portions of the spur is distinctly separate
to varying degrees from the medial margin of the fifth upper
pharyngeal tooth plate.
Juveniles of Curimata cyprinoides have a medial spur of the
fourth epibranchial with an overall triangular form (Figure 7).
At that stage in ontogeny the spur has a moderate area of
attachment basally to the articular process of the bone
(AVPE4), a condition similar to that in adults of many
non-Curimatopsis curimatids. During development the rela-
tively narrow basal portion of the medial spur characteristic
of juveniles of Curimata expands into a broad triangular sheet
of bone extending ventrally to the margin of the articular
surface between the fourth epibranchial and the fifth upper
AVPE4
FIGURE 7.?Curimata cyprinoides, USNM 231433, juvenile, fourth epibran-
chial and fifth upper pharyngeal tooth plate, right side, medial view, anterior
to left, both elements cartilaginous.
phargyneal tooth plate (Figure 8). As a consequence the basal
portion of the spur in these taxa is continuous along an
extensive vertical line with the medial surface of the
ventralmost portion of the body of the fourth epibranchial. An
additional consequence of that ontogenetic expansion is that
specimens of species of Curimata of greater than approxi-
mately 40 mm SL have the ventral margin of the medial spur
of the fourth epibranchial (MBS) in close proximity to the
medial edge of the fifth upper pharyngeal tooth plate (UP5).
These two elements are furthermore relatively tightly attached
by a connective tissue sheet. The dorsal margin of the E4 spur
in species of Curimata thickens progressively as the relative
overall size of the spur increases; a process that ultimately
forms a distinct undercut median shelf at the dorsal margin of
the spur in the adults. The expanded median bony spur on E4
DCE
DFPB3
PB2
PBr
MPE3
MBS
AVPE4
FIGURE 8.?Curimata cyprinoides, USNM 267964, dorsal portion of gill arches, right side, medial view, anterior
to left (dense patterned stippling represents cartilage).
NUMBER 471 17
DCE
MBS
AVPE4 VPE4
FIGURE 9.?Psectrogaster curviventris, USNM 243221, dorsal portion of gill arches, right side, medial view,
anterior to left (dense patterned stippling represents cartilage).
DCE
PB
AVPE4
UP5
FIGURE 10.?Potamorhina laticeps, USNM 121325, dorsal portion of gill arches, right side, medial view, anterior
to left (dense patterned stippling represents cartilage).
DCE,
AVPE4 MBS VPE4 UPC
FIGURE 11.?Steindachnerina hypostoma, USNM 167802, dorsal portion of gill arches, right side, medial view,
anterior to left (dense patterned stippling represents cartilage).
18 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
in adults of Curimata is also distinctive in its pronounced
fenestration (Figure 8). Neither the dorsomedian shelf nor the
fenestration of the fourth epibranchial spur are found in other
curimatids.
The ontogenetic restructuring of the E4 spur of Curimata
from a condition comparable to that generalized for spur-
bearing curimatids to the triangular plate characteristic of the
genus is congruent with a hypothesis of the derived nature of
that ossification (SYNAPOMORPHY 78). Similarly the develop-
ment of a distinct dorsomedial shelf on the medial spur of the
fourth epibranchial, and the fenestration of the spur are
considered synapomorphous for the genus (SYNAPOMORPHY
79). Those ontogenetically derived polarity hypotheses are
furthermore in agreement with data from outgroup analysis
which also indicates that the possession of such elaborations
of the fourth epibranchial spur are unique to, and represent
synapomorphies for, the species of Curimata.
Apomorphous alterations of the dorsal portions of the fourth
epibranchial, both ossified and cartilaginous, are notable within
the Curimatidae. The form and orientation of the dorsal process
of the fourth epibranchial generalized for, and hypothesized
as primitive, within the Characiformes is a transversely
flattened plate with a vertically or nearly vertically oriented
central axis (Figure 4; Vari, 1983, fig. 15). Members of the
Curimatidae (Figures 5, 7-11) and Prochilodontidae (Figure
6) are characterized by having that primary axis of the dorsal
process of E4 reoriented anterodorsally to some degree. That
reorientation, apomorphous for the two-family clade (SYNAPO-
MORPHY 1), is least pronounced in the Prochilodontidae, but
is more developed in some groups of curimatids as will be
discussed below. The distal cartilaginous portion of the fourth
epibranchial also varies in morphology within the Curimatidae.
Outgroup studies indicate that the cartilage capping the dorsal
process of E4 is primitively a relatively small body as typified
by the condition in Xenocharax Gunther (Figure 4) and Brycon
Muller and Troschel (Vari, 1983, fig. 15). That cartilage body
on E4 in the Prochilodontidae is somewhat, but not markedly,
realigned anterodorsally relative to the condition in most
characiforms. That shift and the comparable condition and
more pronounced realignments within the Curimatidae together
represent a synapomorphy for the Curimatidae and Prochi-
lodontidae (SYNAPOMORPHY 2).
Within the Curimatidae further modifications are found both
in the overall form of the dorsal process of E4 and in the
alignment and shape of its associated distal cartilage. The
species of Curimatopsis are characterized by a distal cartilage
of the fourth epibranchial that is similar in form to that in the
Prochilodontidae; however, that genus has the primitively
dorsal process of the fourth epibranchial more anterodorsally
directed than in prochilodontids (Figure 5). That apomorphous
antcrodorsal reorientation of the dorsal process of E4 is more
pronounced in Potamorhina, Curimata, Psectrogaster, Steinaacti-
ne rina, Pseudocurimata, Curimatella, and Cyphocharax.
Those taxa have the primitively dorsal process of E4 more
anteriorly directed (Figures 8-11) than in either the Prochi-
lodontidae (Figure 6) or Curimatopsis (Figure 5). As a
consequence of this apomorphous reorientation, the trans-
versely flattened cartilaginous terminal portion of the fourth
epibranchial (DCE^ constitutes the anterior limit of that
element in non-Curimatopsis curimatids contrary to its dorsal
location on the fourth epibranchial that is typical for
characiforms. Congruent with the reorientation of the dorsal
portion of E4 is the anterior expansion of the distal cartilage
on that element (DCE4) into an elongate plate. In different
phyletic lineages within the Curimatidae that plate extends
anteriorly to varying degrees dorsal of the fourth infrapharyn-
gobranchial and above the portions of the other gill arch
elements located along the median margin of the dorsal half
of the branchial complex (Figures 8, 9, 10, 12, 13).
The absence of comparable modifications of the primitively
dorsal process of the fourth epibranchial and its associated
terminal cartilage in characiform outgroups is congruent with
the hypothesis that these alterations are shared derived
characters for the Curimatidae and its subunits at two levels
of generality. The moderate degree of anterodorsal alignment
of the primary axis of the process in Curimatopsis, and the
further derived more anterior alignment of the process in other
curimatids together constitute a synapomorphy for the entire
family (SYNAPOMORPHY 27). The pronounced anterior realign-
ment, the more derived condition, in Potamorhina, Curimata,
Psectrogaster, Steindachnerina, Pseudo curimata, Curimatella,
and Cyphocharax is a derived synapomorphy of lesser
generality within the family (SYNAPOMORPHY 52). The latter
organization of the ossified and cartilaginous portions of E4 is
considered the condition from which further modifications of
the fourth epibranchial in several less inclusive subunits of the
Curimatidae are derived.
Among the species of Potamorhina the anteriorly reoriented
dorsal process of the fourth epibranchial common to
non-Curimatopsis curimatids is apomorphously further
lengthened to a significant degree. This elaboration of the
process results in a horizontally markedly elongate, vertical
wall that extends medially along the anterolateral portion of
the muscular epibranchial organ (Figure 10). The transversely-
flattened cartilage (DCE4) that caps the primitively dorsal, now
anterior, portion of the bone is congruently significantly
lengthened longitudinally in Potamorhina, and extends for-
ward to contact the dorsal margin of the cartilage that caps the
uncinate process of the third epibranchial (DCE3). A distinct
separation of these two cartilage bodies typifies other examined
characiforms both outside of the Curimatidae (e.g., Xenocharax
spilurus and Prochilodus nigricans, Figures 4, 6), and within
the family (Figures 5, 8, 9, 11). The anterior lengthening of the
dorsal process of the fourth epibranchial along with the close
association of the cartilages of the third and fourth epibranchi-
als in Potamorhina is consequently hypothesized to be
synapomorphous for the members of the genus (SYNAPOMOR-
PHIES 64 and 65).
NUMBER 471 19
The overall elongation of the fourth epibranchial in
Potamorhina is paralleled by the anterior lengthening of the
fifth epibranchial in the members of the genus. The fifth
epibranchial (E5) in most characiforms is typically a small
cartilage at the posterior corner of the fourth epibranchial lying
immediately posterodorsal of the junction of that element and
the fourth ceratobranchial (e.g., Xenocharax spilurus, Figure
4). Both prochilodontids and curimatids have the cartilaginous
E5 expanded into a relatively large mass incorporated into the
lateral wall of the large epibranchial organ. As noted by Vari
(1983) the fifth epibranchial in those families is also
noteworthy in its attachment to the posterodorsal margin of the
fourth epibranchial, with a resultant encirclement by those
elements of the fifth efferent branchial artery. Together this
complex is considered a synapomorphy for the Curimatidae
and Prochilodontidae (SYNAPOMORPHY 2). Within the Curimat-
idae the fifth epibranchial in Potamorhina (E5, Figure 10) is
further lengthened beyond the condition in prochilodontids and
other curimatids into a long tapering body extending anteriorly
to the level of the middle or posterior portion of the fourth
infrapharyngobranchial, a synapomorphy unique to members
of that genus (SYNAPOMORPHY 70).
All of the diverse groups of characiforms examined in the
present study have a distinct cord-like ligament (LIG) that
extends posteriorly from the rear margin of the uncinate
process of the third epibranchial (E3) to an attachment on the
fourth epibranchial (E4) (e.g., Xenocharax spilurus and
Prochilodus nigricans, Figures 4, 6). This inter-epibranchial
ligament attaches to E4 at the anteriormost limit of the distinct
ridge that extends along the lateral surface of the main body
of the fourth epibranchial. In most characiforms that ridge on
the fourth epibranchial terminates anteriorly just dorsal of the
vertical plane through the area of contact of the fourth
infrapharyngobranchial (PB^ and the fourth epibranchial
(Figure 12). The exact function of this inter-epibranchial
ligament is uncertain, but it apparently limits the degree of
flexure of the fourth epibranchial relative to the more anterior
arches of the branchial apparatus, and is one of a series of
comparable ligaments in the dorsal portion of the gill arches.
Shorter, possibly serially homologous, ligaments also extend
between the first epibranchial and the second infrapharyn-
gobranchial, and between the second epibranchial and the third
infrapharyngobranchial.
The five species of Curimatopsis have the ligament between
the third and fourth epibanchials attached onto the relatively
short ridge located on the anterolateral surface of the fourth
epibranchial, an attachment comparable to the condition
described above for characiform outgroups (Figure 12). The
species of Potamorhina, Curimata, Psectrogaster, Steindach-
nerina, Pseudocurimata, Curimatella, and Cyphocharax rather
than retaining a direct attachment of the ligament onto the
main body of the fourth epibranchial have that band inserting
onto a bony spur-like process (AS) that extends forward from
the anterior margin of the main body of E4 (Figures 8, 10, 11,
13, 14). The base of this bony spur corresponds in position to
the region of E4 where the E3-E4 ligament attaches in other
characiforms. The spur, in turn, is longitudinally aligned along
the axis of the ligament. Although the degree of development
of the spur varies among those curimatids characterized by
that process, the common possession of such an anterior
extension of E4 associated with the E3-E4 ligament is
hypothesized to be a synapomorphy for the genera Pota-
morhina, Curimata, Psectrogaster, Steindachnerina, Pseu-
docurimata, Curimatella, and Cyphocharax given the absence
of such an elaboration in examined outgroups (SYNAPOMOR-
PHY 53).
Finally the overall form of the ventral portion of the fourth
epibranchial is modified in all species of Curimata. In other
curimatids, the portion of the fourth epibranchial that
articulates with and supports UP5 is situated directly ventral
of the main body of E4, and is in line with the vertical axis
through the dorsal process of that element In species of
Curimata, alternatively, the ventral portion of E4 is further
shifted medially which results in an associated move towards
DCE
DCE
FIGURE 12.?Curimatopsis microlepis, USNM 268867, dorsal portion of third
and fourth gill arches, left side, dorsolateral view, anterior to left (dense
patterned stippling represents cartilage).
DCE4
DCE
FIGURE 13.?Curimatella alburna, MZUSP 6309, dorsal portion of third and
fourth gill arches, left side, dorsolateral view, anterior to left (dense patterned
stippling represents cartilage).
20 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
the median plane of the attached fifth upper pharyngeal tooth
plate (UP5). As a consequence of this restructuring the complex
consisting of the ventral portion of E4 and the associated UP5
in Curimata is located distinctly medial of the primary vertical
plane through the dorsal section of the branchial apparatus
(Figure 14) (SYNAPOMORPHY 80).
Fifth Upper Pharyngeal Tooth Plate (UP5)
The ventral surface of the fifth upper pharyngeal tooth plate
of most characiforms is typically relatively flat and dentigerous
(e.g., Brycon falcatus. Figure 17, and Xenocharax spilurus,
Figure 4). That tooth plate extends posteriorly a limited
horizontal distance, from the posterior limit of the fourth upper
pharyngeal tooth plate (UP4) to the middle of the horizontal
extent of the fourth epibranchial. The broad range in the form
of the fifth upper pharyngeal tooth plate (UP5) within the
Curimatidae serves to define a number of subassemblages of
the family. All members of the Curimatidae have the amount
of dentition on the fifth upper pharyngeal tooth plate reduced,
even in juveniles, relative to the condition in outgroups
(Figures 5,7-11), a derived feature (SYNAPOMORPHY 5). Both
the fifth upper pharyngeal tooth plate and the associated
dentition undergo further modifications within the family.
The shortest and least elaborate version of UP5 occurs in the
species of Curimatopsis. The tooth plate in that genus has a
distinct patch of teeth on its ventral surface and a short
dorsomedial shaft posteriorly (Figures 5, 15). The overall
longitudinal dimension of UP5 relative to E4 is only somewhat
longer than the state of that element typical of prochilodontids
(e.g., Prochilodus nigricans, Figure 16) and examined
characiform outgroups (e.g., Brycon falcatus, Figure 17). In
the species of Potamorhina, Curimata, Psectrogaster, Steindach-
nerina, Pseudocurimata, Curimatella and Cyphocharax the
fifth upper pharyngeal tooth plate is more elongate longitudi-
E 3
nally (Figures 7-11) with the relatively short posterior shaft
of the element characteristic of Curimatopsis lengthened to
various degrees. The shape of the lengthened posterior portion
of UP5 mirrors the form of the region of the fifth
ceratobranchial which it approximates when the dorsal and
ventral portions of the gill arches are in contact. Although a
number of genera of non-Curimatopsis curimatids have a
variable number of teeth on the ventral surface of the fifth
upper pharyngeal tooth plate as juveniles (Figure 7), the adults
of those taxa are characterized by a pronounced reduction, or
more typically an elimination of the dentition associated with
UP5. On the basis of outgroup comparisons both the increase
in the relative length of UP5, and the decrease in the amount
of dentition on that tooth plate are hypothesized to be shared
derived characters uniting Potamorhina, Curimata, Psec-
trogaster, Steindachnerina, Pseudocurimata, Curimatella, and
Cyphocharax (SYNAPOMORPHY 54).
The apomorphous elongate form of UP5 typical for
non-Curimatopsis curimatids is further modified in various
ways in some less inclusive subunits of the Curimatidae. Only
one of these modifications is, however, derived for a group
herein recognized at the generic level and thus appropriate for
this discussion. Species of the genus Psectrogaster have an
extremely elongate fifth upper pharyngeal tooth plate which
extends caudally beyond the transverse plane through the
posterior limit of the ossified portions of the gill arches (Figure
9). Such a posterior extension of the bone in Psectrogaster
contrasts with the typical curimatid condition in which the fifth
upper pharyngeal tooth plate barely reaches, or more usually
falls distinctly short of, that transverse plane. The distal portion
of UP5 of species of Psectrogaster is also twisted along its
PB PB2
FIGURE 14.?Curimaia cyprinoides, USNM 225619, dorsal portion of gill
arches, left side, ventral view, anterior to left (dense patterned stippling
represents cartilage).
FIGURE 15.?Curimatopsis microlepis, MZUSP 21053, dorsal portion of gill
arches, left side, ventral view, anterior to left (dense patterned stippling
represents cartilage).
NUMBER 471 21
longitudinal axis, and is transversely expanded distally into
an oar-shaped process; an apomorphous restructuring of the
element again unknown elsewhere in the family. In light of the
unique nature of the lengthening of UP5 and its reconfiguration,
the overall form of the bone is considered a derived character
synapomorphous for the species of Psectrogaster (SYNAPOMOR-
PHY 87).
Fourth Upper Pharyngeal Tooth Plate (UP4) and
Fourth Infrapharyngobranchial (PB4)
In both the Curimatidae (Figure 15) and Prochilodontidae
(Figure 16) an edentulous ossification envelopes most of the
ventral surface and portions of the lateral margins of the
cartilaginous fourth infrapharyngobranchial (PB4). The homol-
ogy of this ossification is somewhat problematical. It may
represent a new perichondral element, with the fourth upper
pharyngeal tooth plate (UP4) usually attached to the ventral
surface of the fourth infrapharyngobranchial being absent.
Alternatively, the ossification along the ventral and lateral
surfaces of PB4 in the Curimatidae and Prochilodontidae may
represent a restructured edentulous fourth upper pharyngeal
tooth plate. The latter homology is tentatively proposed since
it requires only one ad hoc hypothesis, the modification in the
overall structure of UP4. The alternative hypothesis, that the
ossification represents a new perichondral element, necessitates
two assumptions, the loss of UP4 and the independent
development of a new ossification on PB4, a more complex
scenario. A more definitive statement on the homology of the
ossification on the ventral portion of PB4 would necessitate
presently unavailable ontogenetic information.
A definitive statement on the homology of the element is
not crucial in terms of its phylogenetic significance. If the
ossification is the fourth upper pharyngeal tooth plate it would,
as was noted by Vari (1983:18), represent a pronounced
restructuring of the plesiomorphous plate-like UP4 (Figure 17)
that typically contacts only the ventral surface of PB4 in most
characiforms. Similarly a new perichondral ossification on the
ventral surface of PB4 and the loss of UP4 would also be
derived. The expanded curved ossification on PB4 was
consequently considered derived by Vari (1983) and a
synapomorphy for the Curimatidae and Prochilodontidae
(SYNAPOMORPHY 4). Subsequent outgroup analysis has shown
that the overall profile of the resultant UP4-PB4 bone-cartilage
complex in those families is also phylogenetically informative.
The typical, and hypothesized plesiomorphous, form of the
fourth infrapharyngobranchial in characiforms has a relatively
straight or moderately convex lateral margin when viewed from
a ventral view. That form of PB4 is common to the majority
of characiforms including several diverse families: Cithar-
inidae (Citharinus Cuvier); Distichodontidae (Neolebias Steindach-
ner (Vari, 1979:305), Mesoborus Pellegrin, and Xenocharax);
Characidae, including genera from the Neotropics (Agoniates
Miiller and Troschel, Brycon (Figure 17), Clupeacharax,
Exodon Muller and Troschel, Hollandichthys Eigenmann) and
the Old World (Alestes Muller and Troschel, Hydrocynus
Cuvier (Brewster, 1986, fig. 14), Phenacogrammus); Chi-
lodontidae (Caenotropus Gunther (see Vari, 1983, figs. 18,
19)); Hemiodontidae {Hemiodopsis Fowler); and Parodontidae
(Parodon Valenciennes). In all examined members of the
Curimatidae and Prochilodontidae the anterior portion of PB4
and the associated anterior section of UP4 are transversely
expanded laterally which results in the overlap of the
anteromedial surface of the third epibranchial (E3) by the
FIGURE \6.?Prochilodus nigricans, USNM 231438, dorsal portion of gill
arches, left side, ventral view, anterior to left (dense patterned stippling
represents cartilage).
FIGURE 17.?Brycon falcatus, USNM 226161, dorsal portion of gill arches,
left side, ventral view, anterior to left (dense patterned suppling represents
cartilage).
22 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
conjoined PB4 and UP4. As a consequence of this anterior
expansion, the PB4-UP4 complex appears transversely con-
stricted midway along its longitudinal length (Figures 15, 16)
contrary to the relatively parallel, or outwardly bowed lateral
margin of that complex in other examined characiforms (e.g.,
Bryconfalcatus, Figure 17). Outgroup comparisons have failed
to reveal comparable modifications in other examined characi-
forms, and the form of the PB4-UP4 bone-cartilage complex is
thus considered synapomorphous for the members of the
Curimatidae and Prochilodontidae (SYNAPOMORPHY 12).
The form of the conjoined UP4-PB4 common to curimatids
and prochilodontids is further modified in one generic level
subunit of the Curimatidae. Potamorhina species have both the
fourth infrapharyngobranchial and the associated fourth upper
pharygneal tooth plate significantly lengthened longitudinally
(Figure 10). That adaptation is considered derived (S YNAPOMOR-
PHY 66) with respect to the relatively shorter form of these
elements in other curimatids and the families that constitute
the nested sister groups to the Curimatidae .
Third Epibranchial (E^
The dorsal portion of the third branchial arch consists of two
bones, the third epibranchial (E3) and third infrapharyngobran-
chial (PB3) along with their associated cartilages. Derived
modifications of those elements unite the Prochilodontidae and
Curimatidae, and define lineages within the latter family. The
apomorphous modifications of the third epibranchial will be
discussed first.
The generalized form of the third epibranchial among
characiforms has a cartilage capped uncinate process (UN)
whose distal portion (DCE3) terminates both distinctly lateral
of the median plane of the gill arches, and posterior of the
junction of the third infrapharyngobranchial with the third
epibranchial and fourth infrapharyngobranchial (Figures 4,
18). All members of the Curimatidae and Prochilodontidae
have the uncinate process of the third epibranchial markedly
modified. In those two families the uncinate process is oriented
anteriorly (Figure 19) rather than retaining the lateral or slightly
anterolateral alignment found in other characiforms (Figure
18). As a consequence of this realignment the uncinate process
with its associated distal cartilage overlays the vertical plane
through the articulation between PB3 and E3 rather than
terminating distinctly posterior of that articulation. The
uncinate process additionally more closely approaches the
median plane of the arches than is typical in characiform
outgroups. Correlated with these changes is the approximation
with the uncinate process to the midventral basal portion of the
neurocranium to which it is attached by a ligamentous band.
Together these hypothesized derived modifications are consid-
ered a synapomorphy for the Curimatidae and Prochilodontidae
(SYNAPOMORPHY 13).
The anterior portion of the main shaft of E3 in characiform
outgroups is capped anteriorly by a relatively large articular
cartilage that contacts the lateral half of the posterior margin
of the cartilaginous third infrapharyngobranchial (PB3). Me-
dially that cartilage cap of the third epibranchial (ASE3, Figure
18) slightly overlies and abuts the dorsolateral surface of the
fourth infrapharyngobranchial (PB^, but is distinctly separated
from the midsagittal plane. The anterior portion of the main
shaft of E3 in the Curimatidae and Prochilodontidae is
significantly modified relative to that generalized bauplan.
ASE
DCE
FIGURE \%.?Bryconfalcatus, USNM 226161, fourth infrapharyngobranchial,
posterior portion of third infrapharyngobranchial, and anterior portion of third
epibranchial, right side, dorsal view, anterior to left (dense patterned stippling
represents cartilage).
FIGURE 19.?Curimata cisandina, USNM 229171, fourth infrapharyngobran-
chial, fourth upper pharyngeal tooth plate, posterior portion of third
infrapharyngobranchial, and third epibranchial, right side, dorsal view, anterior
to left (dense patterned stippling represents cartilage).
NUMBER 471 23
Rather than terminating as a single articular surface along the
posterolateral margin of the third infrapharyngobranchial, the
articular surface of the third epibranchial in the Curimatidae
and Prochilodontidae is divided into two subsections. The
anterolateral margin of E3 retains contact with the cartilage
along the posterolateral portion of PB3 (LPE3, Figure 19). In
contrast, the relatively small medial extension of the antero-
lateral portion of the third epibrancial that characteristically
extends over the dorsolateral surface of the fourth infrapharyn-
gobranchial of most characiforms is totally altered in these two
families. Curimatids and prochilodontids have the medial
extension expanded into a relatively thick ossified process
(MPE3, Figure 19) which extends over the anterodorsal portion
of the fourth infrapharyngobranchial and terminates medially
in a relatively large cartilage cap. This medially expanded
process extends along and is surrounded posteroventrally by a
matching groove in the anterodorsal surface of the cartilaginous
fourth infrapharyngobranchial (PB^). As a consequence the
medial process of the third epibranchial in both families is
visible in a medial view of the dorsal portions of the gill arches
(Figures 5, 6, 8-11). The lateral expansion and ossification of
the medial process of the third epibranchial, and associated
reconfiguration of the proximate portion of PB4 in the
Curimatidae and Prochilodontidae are considered derived
relative to the condition of these elements in other characi-
forms, and together represent a synapomorphy for those
families (SYNAPOMORPHY 14).
A medial elongation of the anterior portion of PB3 also
occurs among characiforms in the Anostomidae (Vari, 1983,
fig. 20), Chilodontidae (Vari, 1983, fig. 19), Hemiodontidae,
and Parodontidae. None of the cited families has the third
epibranchial terminating anteriorly in a large medial process
comparable to that of the Curimatidae and Prochilodontidae.
With the exception of the Chilodontidae (Chilodus Miiller and
Troschel, Caenotropus) and a single genus in the Hemiodon-
tidae (Bivibranchia Eigenmann, sensu Vari, 1985), the
cartilage along the anteromedial portions of the articular
surface on the third epibranchial in these families is continuous
with the lateral portion of that cartilage contacting the fourth
infrapharyngobranchial. Such continuity contrasts with the
subdivision of that cartilage cap into two independent masses
in the Curimatidae and Prochilodontidae. As noted, the
Chilodontidae has two distinct distal cartilages on the anterior
portion of E3 (e.g., Caenotropus, Vari, 1983, fig. 19). The
medial section of that portion of the bone, however, consists
of a slender shaft approximating but not reaching the medial
plane, a significantly different condition than that found in the
Curimatidae and Prochilodontidae.
Furthermore the Chilodontidae shares numerous synapomor-
phies with, and is evidently most closely related to the
Anostomidae (Vari, 1983:50) which lacks comparable derived
restructurings of the anterior portion of the third epibranchial.
The medial process of the third epibranchial in Bivibranchia
more closely approximates the form of the structure in the
Curimatidae and Prochilodontidae, but differs from the
condition in those families in details of its site of origin and
overall morphology. Moreover, Argonectes Bohlke and Myers,
which is evidently the closest relative of Bivibranchia (Roberts,
1974), together with all the other members of the Hemiodon-
tidae lack such a distinct medial process on E3. The noted
differences in the structure of the third epibranchial in the
Curimatidae and Prochilodontidae versus those in the Chi-
lodontidae and Bivibranchia, in conjunction with the data on
phylogenetic relationships of those taxa (Vari, 1983; Roberts,
1974) make it simplest to assume that the modifications in the
Chilodontidae and Bivibranchia are non-homologous with the
described alterations of the third epibranchial hypothesized
synapomorphous for the Curimatidae and Prochilodontidae.
All members of the Curimatidae have at least a slight ridge
along the ventral surface of E3 extending in line with the
longitudinal axis of the element. This ridge is further elaborated
in two subunits of the family. A well developed ridge is
characteristic of all members of Psectrogaster (SYNAPOMOR-
PHY 88). Similarly a very strong ventral ridge on E3 is shared
by a clade consisting of three nominal species within Curimata
(aspera, simulata, and cerasina); an elaboration which in those
species parallels the ridge on the opposing dorsal surface of
the third ceratobranchial of the ventral portion of the branchial
basket. The common possession of the ventral ridge in
Psectrogaster and a subunit of Curimata is considered
homoplasious in the overall most parsimonious scheme of
relationships within the family, although synapomorphous for
each of those clades (see "Synapomorphy List and Phylo-
genetic Reconstruction" and "Convergent Characters").
Third Infrapharyngobranchial (PB3)
The third infrapharyngobranchial (PB3) is typically a
flattened or slightly dorsally concave, longitudinally elongate
triangular bone in characiforms. As a consequence of its
flattened form the ossification lies entirely or nearly entirely
in the horizontal plane through the ventral surface of the upper
portions of the gill arches. The lateral portion of PB3 is
consequently not visible or only slightly visible in a medial
view of that complex (e.g., Xenocharax spilurus. Figure 4).
Such a flattened PB3 characterizes the Anostomidae and
Chilodontidae which together constitute the sister clade to the
lineage formed by the Prochilodontidae and Curimatidae. The
Prochilodontidae and Curimatidae rather than retaining the
flattened PB3 of other characiforms have the posterolateral
portion of the ossification restructured into a vertically or near
vertically aligned flange that extends distinctly dorsal of the
horizontal plane through the ventral main portion of the
ossification. This restructuring results in the lateral portion of
the bone being readily visible in a median view of the gill
arches in both curimatids and prochilodontids (DFPB3, Figures
5,6, 8-11). The unique nature of this modification of the third
pharyngobranchial in those two families is indicative of the
24 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
derived nature of the character and is another derived attribute
supporting the hypothesis of the monophyly of the clade
formed by the Curimatidae and Prochilodontidae (S YNAPOMOR-
PHY 15).
The association of the third infrapharyngobrancial (PB3)
with its respective epibranchial (E3) and of PB3 with the fourth
infrapharyngobranchial (PB4) varies within the family Curi-
matidae. The form of PB3 typical for characiforms has a
continuous articular cartilage along its posterior margin. The
medial portion of the posterior PB3 cartilage contacts the
anterior margin of the cartilaginous PB4, and the lateral portion
of that cartilage of PB3 abuts the articular cartilage that caps
the anterior section of E3 (Figure 16,17). In most characiforms
the joint of PB3 with PB4 and E3 is either transversely straight
(Figure 16) or at a slight angle (Figure 17), but there is no
pronounced overlap of PB3 lateral to the ossified portion of
E3. All Curimatopsis species (Figure 15), in contrast, have the
joint between the third infrapharyngobranchial and third
epibranchial distinctly oblique relative to a transverse plane
through the branchial basket. As a consequence of the
restructuring the posterolateral portion of the third infrapharyn-
gobranchial apomorphously extends posteriorly into the region
primitively occupied by the anterolateral portion of the third
epibranchial (compare Figures 15-17), a derived condition
(SYNAPOMORPHY 43). The overlap between these elements
found in some members of the Old World characiform family
Distichodontidae (e.g., Mesoborus) is a function of the
posterolateral development of a process of PB3 that extends
lateral of the anterolateral edge of the ossified portion of E3
rather than a consequence of the realignment of the PB3-E3
joint, and thus is non-homologous with the condition in
Curimatopsis.
As noted, it appears that a continuous cartilage cap along
the posterior margin of the third infrapharygobranchial is the
primitive condition of that element in characiforms. Such an
articular surface is common to the Prochilodontidae (Figure
16), New and Old World members of the Characidae (e.g.,
Brycon, Figure 17), the Hemiodontidae, Hepsetidae, Cithar-
inidae, Distichodontidae, and other characiform outgroups
examined. Such a continuous cartilage cap on PB3 is also
characteristic of all members of Curimatopsis (Figure 15) in
the Curimatidae. In the species of Potamorhina, Curimata,
Psectrogaster, Steindachnerina, Pseudocurimata, Curimatella,
and Cyphocharax this hypothesized plesiomorphic single
cartilage mass on PB3 is subdivided sagitally (SYNAPOMORPHY
55). The medial portion of the cartilage mass of the third
infrapharyngobranchial articulates with the anterior surface of
the fourth infrapharygobranchial. The lateral section of the
cartilage, located along the dorsal flange of the third
infrapharyngobranchial (DFPB3), contacts the anterior articular
cartilage of E3 (Figures 8, 10, 11). A distinct ossified region
along the posterior border of the third infrapharyngobranchial
separates these two cartilage masses.
Although hypothesized to be a derived character common
to all genera other than Curimatopsis within the Curimatidae,
a subdivision of the posterior articular cartilage on the third
infrapharyngobranchial is not unique to that assemblage among
characiforms. The transversely widened PB3 of the Chilodon-
tidae and Anostomidae demonstrates a similar fission of the
posterior cartilage mass (see Vari, 1983, figs. 18, 19). Within
the Hemiodontidae, a comparable transversely divided articular
surface is limited to Bivibranchia (sensu Vari, 1985). The
overall hypothesis of relationships for the Curimatidae,
Prochilodontidae, Anostomidae, and Chilodontidae (Vari,
1983) indicates that the common presence of a subdivided PB3
cartilage in the majority of the genera in the Curimatidae, and
in the clade consisting of the Anostomidae and Chilodontidae
may represent a homoplasious independent acquisition of that
derived feature in those two lineages. An alternative hypothesis
is its loss in the ancestors of the Prochilodontidae and in the
lineage consisting of Curimatopsis; or by the fusion of those
cartilage bodies in the common ancestor of Curimatidae and
Prochilodontidae and the subsequent fission of the cartilage in
non-Curimatopsis curimatids. The hypothesis of the independ-
ent acquisition of a subdivided cartilage in the clade formedby
the Anostomidae and Chilodontidae, and in non-Curimatopsis
curimatids proposed in this study is more parsimonious by one
step than the cited possible alternative ad hoc hypotheses.
Thus, the possession of two discrete articular cartilages on the
posterior of PB3 is considered a synapomorphy for Pota-
morhina, Curimata, Psectrogaster, Steindachnerina, Pseu-
docurimata, Curimatella, and Cyphocharax.
The hypothesis of the independent acquisition of a
subdivided PB3 articular cartilage within and outside of the
Curimatidae can also be applied from a functional basis. In the
Chilodontidae and Anostomidae the subdivision of the articular
cartilage is apparently correlated with the pronounced trans-
verse widening of the third infrapharyngobranchial. Such a
transverse expansion of the third infrapharyngobranchial is not
found in Potamorhina, Curimata, Psectrogaster, Steindach-
nerina, Pseudocurimata, Curimatella, and Cyphocharax that
rather demonstrate a dorsal expansion of the lateral region of
the bone. The common possession of a subdivided posterior
articular cartilage in non-Curimatopsis curimatids and Bivi-
branchia is similarly considered homoplasious. Available data
indicates that Bivibranchia is a highly modified member of the
Hemiodontidae that is evidently most closely related to
Argonectes (Roberts, 1974:432). Neither Argonectes, the
presumed sister group to Bivibranchia, nor any other of the
examined hemiodontid genera (Anodus, Hemiodopsis, Hemi-
odus Miiller and Troschel, Micromischodus Roberts) have the
transverse development of the medial portion of PB3 typical
of Bivibranchia, nor do they have the transverse subdivision
of the posterior articular cartilage on that element that
characterizes that genus. It is thus simplest to consider the
presence of a subdivided PB3 articular surface in Bivibranchia
as being independently acquired and thus not homologoous to
the similar modifications in non-Curimatopsis curimatids.
NUMBER 471 25
Second Branchial Arch
The second epibranchial (E2) and second infrapharyngobran-
chial (PBj) of characiforms typically articulate along a plane
nearly transverse to the main longitudinal axis of the branchial
complex. The second epibranchial ranges from rounded to
somewhat V-shaped in cross section, but with its ventral
margin rounded in either case. The second infrapharyngobran-
chial is flanked anterolaterally by the anteromedial portion of
the first epibranchial, and along its medial surface contacts the
lateral surface of the anterior half of the third infrapharyn-
gobranchial. The contact between these elements is strength-
ened by ligamentous connections, but not via elaborations of
the ossified portion of PB2.
The complex consisting of the second epibranchial and the
second infrapharyngobranchial exhibits within the Curimatidae
some modifications similar to those that occur on the third
branchial arch in that family. The angled area of contact noted
above for the third epibranchial and third infrapharyngobran-
chial in Curimatopsis is paralleled and indeed carried even
further on the second gill arch (Figure 15). All Curimatopsis
species have a significant reduction of the anterolateral margin
of the second epibranchial. The lateral margin of PB2 in the
species of that genus is, in turn, expanded laterally into the
region occupied in other characiforms and curimatids by the
anterolateral portion of E2. This significantly increases the size
of PB2 and results in a unique transversely angled articulation
between the second epibranchial and second infrapharyn-
gobranchial. These conditions are together considered a
snyapomorphy for the genus (SYNAPOMORPHY 44).
All species of Psectrogaster have longitudinally aligned
ridges along the ventral surface of the second epibranchial
comparable to those of the third epibranchial. These ridges,
which contrast with the ventrally unelaborated second epibran-
chial generalized for characiforms, are another derived
character hypothesized synapomorphous for the members of
the genus (SYNAPOMORPHY 89). Similar ridges are found in a
subunit of Curimata (mivartii, cerasina, aspera, simulata,
alleni, cisandina, and two undescribed species). As in the case
of the ridge on the third epibranchial, the most parsimonious
hypothesis of relationships within the Curimatidae (see
"Synapomorphy List and Phylogenetic Reconstruction" and
"Convergent Characters") indicates that the ridges in Psec-
trogaster and the cited subunit of Curimata represent
independent acquisitions of those elaborations.
The medialmost element of the dorsal portion of the second
branchial arch, the second infrapharyngobranchial (PB2) also
demonstrates various derived modifications. The posterolateral
elaboration of the element associated with the expansion of the
articular surface between the second infrapharyngobranchial
and second epibranchial in Curimatopsis was noted above. In
the species of Steindachnerina the anteromedial surface of the
third infrapharyngobranchial fits into a depression on the lateral
surface of the second infrapharyngobranchial (PB2) (Figure
DLF
VLF
FIGURE 20.?Steindachnerina hypostoma, USNM 167802, second infrapharyn-
gobranchial, right side, posterior view, medial to left (dense patterned stippling
represents cartilage).
20). Ventrally PB2 has an ossified shelf (VLF) extending along
the ventral margin of PB3 and dorsally there is a vertical plate
(DLF) extending along the medial surface of that element.
These processes are moderately developed in the nominal
species and subspecies S. argentea, S. conspersa, S. bimacu-
lata, S. trachysteta, S. bimaculata sialis, S. pterostigma, S.
semiornata, S. melanira, S. leucisca leucisca, S. leucisca
bolivae, and S. binotata, and more elaborately developed in the
remaining members of the genus. The possession of such
processes whether moderately or well developed is considered
a synapomorphy for the genus (SYNAPOMORPHY 97).
Psectrogaster species also have a distinct flange on the
dorsal surface of PB2 (SYNAPOMORPHY 90) comparable to that
in Steindachnerina, particularly those species of Steindachner-
ina with the more elaborate vertical development of the flange.
However Psectrogaster lacks the ventral shelf overlapping the
ventral surface of PB3. Although comparable in overall form,
I consider the similarly placed dorsal processes on the lateral
surface of the second infraphyarngobranchial in Steindachner-
ina and Psectrogaster to be homoplasious within the context
of the overall most parsimonious hypothesis of relationships
(see "Convergent Characters").
First Branchial Arch
The typical characiform first branchial arch has a ventrally
unelaborated first epibranchial (Ej) and a moderate sized first
infrapharyngobranchial (PBj) whose cartilaginous basal sec-
tion is smaller than the ossified portion which typically forms
the dorsal section of that element The dorsally located uncinate
process of the first epibranchial extends from the anteromedial
surface of Ej, having the form of a discrete process terminating
anteriorly in a cartilaginous body that contacts a corresponding
cartilage on the dorsal surface of the uncinate process of the
second infrapharyngobranchial (PB2).
Both of the elements of the first branchial arch, the first
26 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
infrapharyngobranchial and first epibranchial, demonstrate
apomorphous modifications within the Curimatidae. In all
members of Curimata, PBj is either proportionally reduced in
size compared to the moderate sized PBj typical of most
characiforms, or is completely lacking. This progressive
reduction and loss of the first infrapharyngobranchial is
hypothesized as derived given the condition in the outgroups,
and is thus a synapomorphy for the species of Curimata
(SYNAPOMORPHY 81). Steindachnerina species do not demon-
strate any marked changes in the ossified portion of the first
infrapharyngobranchial, but the medial cartilaginous section
of the element is greatly enlarged, being more extensive than
the ossified portion of the first infrapharyngobranchial and
contacting the anterior portions of both the first epibranchial
and second infrapharyngobranchial. This condition represents
an evidently derived state relative to the morphology of those
components common to other curimatids in which the ossified
portion of the first infrapharyngobranchial is the larger, and
wherein the first and second infrapharyngobranchials are
typically not in contact (SYNAPOMORPHY 98).
The typical first epibranchial in characiforms is V-shaped
in cross section and most commonly unelaborated on its ventral
surface. A well-developed ridge on the ventral surface of Ej
is typical for the species of Psectrogaster, a hypothesized
derived condition given the absence of such an elaboration in
characiform outgroups (SYNAPOMORPHY 91). A longitudinal
ridge is also present ventrally on the E{ of Curimata species,
but ranges within that genus in its degree of development. The
well-developed ridge comparable to that of Psectrogaster
occurs only in two species of Curimata and the overall most
parsimonious hypothesis of relationships within the Curimat-
idae indicates that the presence of such a ridge in Curimata is
non-homologous with the ridge characteristic of Psectrogaster
species (see "Synapomorphy List and Phylogenetic Recon-
struction" and "Convergent Characters").
Outgroup studies indicate that the first epibranchial in
characiforms has a small to moderate sized, dorsally located
uncinate process that terminates in a single cartilaginous cap
which contacts a corresponding articular cartilage on the dorsal
surface of the second infrapharyngobranchial. Potamorhina
species, in contrast, have a well developed Et uncinate process
which has two distinct subsections distally, each terminating
in a distal cartilage cap (UN, Figure 21) (SYNAPOMORPHY 71).
The anterior of the two articular surface contacts the matching
cartilage body on the uncinate process that arises from the
anteroventral portion of PB2. The posterior cartilage-capped
articular surface of the uncinate process of Ej extends dorsally
from the main body of the process at an angle of approximately
45? relative to its partner and approximates the neurocranium
dorsally. Two separate articular surfaces on the uncinate
process of Ej also occur in all nominal species of Curimata
other than C. ocellata, C. semitaeniata, C. vittata, C. murieli,
and C. macrops. The subdivided cartilage cap on the Ej
uncinate process of most Curimata species differs, however,
FIGURE 21.?Potamorhina squamoralevis, USNM 243228, first epibranchial,
right side, anterior portion, posterior view, median plane to right (dense
patterned stippling represents cartilage).
from the state in Potamorhina in its overall morphology and
details of its spatial relationships with the second infrapharyn-
gobranchial. Within Curimata it is also possible to follow the
ontogenetic and phylogenetic expansion of the distal portion
of the uncinate process of Ej, and the fission of its articular
surface into two subunits. No internal evidence of such a
subdivision exists for the species for Potamorhina, and it is
possible that the situation in that genus does not represent a
subdivision of the original cartilaginous articular surface, but
rather the de novo development of a second cartilage body on
the uncinate process. Furthermore, the overall hypothesis of
relationships within the Curimatidae indicates that these two
cartilage-capped surfaces on the Ej uncinate process in
Potamorhina and a subunit of Curimata are independently
achieved (see "Synapomorphy List and Phylogenetic Recon-
struction").
Ventral Portion of Branchial Arches
Although not demonstrating such dramatic derived restruc-
turings as are found in the dorsal portions of the gill arches,
the ventral elements of the branchial basket have a number of
modifications pertinent to the question of the monophyly of
the Curimatidae and relative to the relationships of its
constituent genera. In the following discussion, the analysis is
arranged around subunits of the arches that typically consist
of individual elements.
First Basibranchial (BBj)
The generalized condition for characiforms is the possession
of a distinct, typically rod-shaped, medial first basibranchial
situated posterior of the basihyal and anterior of the second
NUMBER 471 27
basibranchial. As noted by Vari (1983:14) prochilodontids
have a small triangular first basibranchial that abuts the anterior
margin of the second basibranchial and which barely extends
to the level of the first hypobranchial (see Vari, 1983, fig. 9).
A recognizable first basibranchial is, in turn, lacking in all
curimatids (Figure 25). This reductive trend is hypothesized
to be apomorphous at two levels of universality. The reduction
of the first basibranchial or the absence of that element as a
discrete ossification are together considered a synapomorphy
for the Prochilodontidae and Curimatidae (SYNAPOMORPHY
6). The absence of an ossified first basibranchial is, in turn,
considered a synapomorphy for the members of the family
(SYNAPOMORPHY 19) (see Vari, 1983:14 for a further
discussion of this complex).
Second Hypobranchial (H2)
The articular cartilaginous surface of the second hypobran-
chial of most characiforms consists of two subunits, a smaller
cartilage body at the anterolateral corner of the ossification and
an extensive cartilage along the medial and posterior portions
of the bone (Figure 22A) . This morphology is found in the
Anostomidae, Chilodontidae, and Prochilodontidae, and among
curimatids also occurs in the genus Curimatopsis. Potamorhina
species retain two separate cartilages anteriorly, but the
separation between them is considerably reduced (Figure 22B).
The genera Curimata, Psectrogaster, Steindachnerina, Pseu-
docurimata, Curimatella, and Cyphocharax have a continuous
cartilage body extending along the anterior and at least part of
the medial surface of the element (Figure 22c) rather than two
separate cartilages. In light of the broad separation anteriorly
of the cartilages on the second hypobranchial in characiform
outgroups, the reduction or elimination of the gap between the
two anterior cartilages is considered a synapomorphy for the
members of Potamorhina, Curimata, Psectrogaster, Steindach-
nerina, Curimatella, and Cyphocharax (SYNAPOMORPHY 56).
FIGURE 22.?Second hypobranchial: A, Curimatopsis microlepis, USNM
268867; B, Potamorhina laliceps, USNM 121325; C, Psectrogaster amazonica,
USNM 261518; and D, Pseudocurimata peruana, USNM 285667. (Right side,
dorsal view, anterior at top, lateral to left, dense patterned stippling represents
cartilage.)
The fusion of the two anterior cartilages into a continuous
body in Curimata, Psectrogaster, Steindachnerina, Pseudocu-
rimata, Curimatella, and Cyphocharax is, in turn, hypothesized
to be a synapomorphy for those taxa (SYNAPOMORPHY 75) (see
also discussion immediately following).
The continuous cartilage along the anterior border and
anterior two-thirds of the medial margin of the second
hypobranchial in that clade undergoes a secondary fission in
the species of Pseudocurimata. In those taxa both the ossified
and cartilaginous portions of the second hypobranchial
proximate to the anterolateral margin of the third basibranchial
are expanded vertically to form a more extensive articular
surface (Figure 22D). Associated with that expansion is the
separation of the continuous articular cartilage characteristic
of proximate sister groups of Pseudocurimata into a broad
anterior section and an elongate medial portion, with the two
cartilages separated by an ossified region at the anteromedial
angle of the bone. The expansion of the medial portion of the
second hypobranchial and the subdivision of the originally
continuous cartilage body along its anterior and medial margins
are together considered a synapomorphy for the species of
Pseudocurimata (SYNAPOMORPHY 101).
As noted above, a subdivided cartilage along the anterior
and medial margins of the second hypobranchial also occurs
in Curimatopsis and Potamorhina, in which that condition was
hypothesized to be plesiomorphous. The forms of the cartilages
on this element in Curimatopsis and Potamorhina on the one
hand and Pseudocurimata on the other differ both in their
relative locations and overall forms. The anterior cartilage in
Curimatopsis and particularly Potamorhina is a small conical
body limited to a process on the anterolateral corner of the
bone. The anterior cartilage of Pseudocurimata, in contrast, is
a narrow body extending over a broad portion of that margin
of the bone. Similarly the gaps between the two cartilages are
located in different regions in the two genera (compare Figure
22A.B with 22D). Futhermore the thickening characteristic of
the second hypobranchial in Pseudocurimata species is absent
in Curimatopsis. These morphological differences lead to a
hypothesis of the non-homology of the two cartilage bodies
on the anterior and medial margins of the second hypobranchial
in Curimatopsis and Potamorhina versus Pseudocurimata, a
hypothesis congruent with the overall most parsimonious
scheme of relationships within the family (Figure 44).
Third Hypobranchial (H3)
The third hypobranchial in characiforms is typically a
flattened ossification that is unelaborated ventrally. The bone
slants anteroventrally lateral to the third basibranchial, and
terminates anteriorly in a rounded to distinctly acute triangular
point. The anterior terminus of the third hypobranchial, be it
an oblique point or an acute angle, serves as the point of
attachment for a ligament that extends to the posteromedial
surface of the second hypobranchial. The just described
28 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
hypothesized plesiomorphous condition of the third hypobran-
chial is notably modified in all species of the Curimatidae. The
members of the family have the anteroventral portion of the
third hypobranchial elaborated into a discrete ventral process
bearing both anterior and posterior extensions (AE and PE;
Figures 23,24). These elaborations, which lie in the connective
tissue complex which parallels the walls of the ventral aorta,
are not typical for characiforms and are hypothesized to be
synapomorphous for the members of the Curimatidae (S YNAPO-
MORPHY 18).
Members of the family Hemiodontidae (Anodus, Hemiodus,
Hemiodopsis (Van, 1982a: 19), Micromischodus, Bivibranchia
[including Atomaster, Vari, 1985:512], and Argonectes) and
one genus in the Chilodontidae (Caenotropus labyrinthicus)
have the anterior portion of H3 more developed ventrally than
is the case in most characiforms. As a consequence the overall
morphology of the anterior section of H3 in those taxa is
somewhat reminiscent of the anterior portion of the ventral
H3 complex of curimatids. Furthermore, two of the taxa,
Anodus and Caenotropus labyrinthicus, also have the vertical
flanges on the ventral surface of the third hypobranchial
expanded in the connective tissue next to the walls of the
ventral aorta, conditions similar to those characteristic of
curimatids. The ventral elaborations and flanges in these genera
BB
AE
IL
differ from the similar processes in curimatids in a number of
attributes. The ligament connecting the second and third
hypobranchials in the Hemiodontidae and Caenotropus arises
from the anterior tip of the anterior process of the third
hypobranchial. That association of the ligament with that
portion of the process is congruent with the hypothesis that the
point of attachment, the elongate anterior point, in those taxa
carried the H2-H3 ligament forward with it during its
phylogenetic expansion anteriorly. Thus the lengthened ante-
rior process of hemiodontids and Caenotropus is a conse-
quence of the greater development of the entire ventral section
of the third hypobranchial. In the Curimatidae, in contrast, the
ligament attaches to the dorsal surface of the ventral process
of the third hypobranchial proximate to the main body of that
element (IL, Figures 23 and 24). The anterior projection of the
ventral process, in turn, extends anteriorly distinctly ventral
of the point of attachment of the ligament on the third
hypobranchial. Thus it appears that the elongate anterior
portion of the bone in the Curimatidae represents an anterior
outgrowth of that portion of the bone ventral of the point of
attachment of the ligament rather than an elongation of the
entire element as is the case in the Hemiodontidae and
Caenotropus.
The cited hemiodontids and chilodontids also differ from
curimatids in the absence of any posterior extension on the
ventral processes of H3. Those differences in the details of the
ventral process of H3 and the overall most parsimonious
hypothesis of phylogenetic relationships of the Curimatidae
to characiform outgroups (Vari, 1983) lead to a hypothesis
that the restructuring of the ventral surface of the third
hypobranchial in the Curimatidae is a synapomorphy for that
family, achieved in a different fashion in the Hemiodontidae
and one genus in the Chilodontidae.
The area of attachment on the third hypobranchial for the
ligament that extends between that element and the second
hypobranchial (Hj) is modified in all Steindachnerina species.
TCBB3 CCH3
FIGURE 23.?Curimata cyprinoides, USNM 267964, third basibranchial, third
hypobranchial, and anterior portion of third ceratobranchials, ventral view,
anterior at top (dense patterned stippling represents cartilage).
FIGURE 24.?Steindachnerina hypostoma, USNM 167802, third basibranchial
and third hypobranchial, right side, lateral view, anterior to right (dense
patterned stippling represents cartilage).
NUMBER 471 29
The connective tissue band between Hj and H3 typically
attaches to the unelaborated anterolateral surface of the vertical
strut of the ventral process of the third hypobranchial, or
attaches to a slightly developed process in that area. In
Steindachnerina species the point of attachment of the ligament
on H3 is rather a distinct process on the anterolateral surface
of the ventral process. This condition is hypothesized to be
derived given the absence of such a process in examined
characiform outgroups and other lineages in the Curimatidae
(IL, Figure 24) (SYNAPOMORPHY 99).
Third Ceratobranchial (C3)
The third ceratobranchial of outgroups to the Curimatidae,
including the Prochilodontidae has its dorsal margin trans-
versely rounded or bearing a slight middorsal ridge extending
along the longitudinal axis of the bone. A discrete longitudinal
ridge along the dorsal surface of the bone characterizes all
Psectrogaster species (SYNAPOMORPHY 92). Within Curimata
a comparable ridge along the dorsal surface of C3 also occurs
in various nominal species (mivartii, aspera, simulata,
cerasina, alleni, and two undescribed species) but differs in its
shorter longitudinal extent along the bone. The ridge in those
Curimata species also has a distinctly convex dorsal margin
which contrasts with the straight edge of the ridge in
Psectrogaster. The presence of a ridge, the derived condition,
is considered to have arisen independently in Psectrogaster
and some Curimata species within the context of the overall
intrafamilial phytogeny arrived at in this study, and in light of
the cited morphological differences.
Fourth Basibranchial (BB4)
The fourth basibranchial in curimatids is a laterally
compressed cartilaginous body without any associated tooth-
plate along its dorsal surface. Only one modification of that
element is pertinent to the current hypothesis of generic level
relationships within the Curimatidae. In Potamorhina the gill
arches are relatively elongate, an overall reconfiguration that
is particularly produced in the fourth basibranchial. In all
members of Potamorhina the fourth basibranchial occupies
approximately two-thirds of the longitudinal extent of the
branchial basket (Figure 25), an apomorphous lengthening
over the condition in other curimatids and a synapomorphy for
the genus (SYNAPOMORPHY 66) (see also Vari, 1984a:5 for a
discussion of this character).
Fourth Ceratobranchial (C4)
Outgroup comparisons reveal that the typical characiform
fourth ceratobranchial although convex in cross section is
unelaborated on its upper surface. All curimatids have various
elaborations of the dorsal surface of the element relative to the
hypothesized simple primitive condition. The species of
BB
FIGURE 25.?Potamorhina squamoralevis, USNM 243228, ventral portion of
gill arches, dorsal view, anterior at top (dense patterned stippling represents
cartilage).
Curimatopsis species have two distinct ridges running along
the dorsomedial and dorsolateral margins of the bone (MR and
LR, Figure 26A). These ridges extend from the wider anterior
portion of the bone posteriorly along the main shaft of C4,
diverging somewhat posteriorly and are separated by a distinct
longitudinal trough. The genera Potamorhina, Curimata,
Psectrogaster, Steindachnerina, Pseudocurimata, Curimatella,
and Cyphocharax, in contrast, have the ridges on the dorsal
surface of the fourth ceratobranchial diverging to some degree
anteriorly rather than being parallel in that region (Figure
26B.C). The ridges in those taxa are also positioned more
towards the medial and lateral margins of the bone than in
Curimatopsis (compare Figure 26A with 26B.C). The anterior
ridge in these genera is furthermore more horizontally aligned
than in Curimatopsis, a shift which results in a discrete bony
pocket along the posteromedial surface of the bone. The lateral
ridge in non-Curimatopsis curimatids, in turn, is rotated
30 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
somewhat (Figure 26B) to distinctly lateral (Figure 26c). The
presence of the ridges in one of these configurations, a
character complex not present in examined characiform
outgroups, is considered a synapomorphy for the species of the
Curimatidae (SYNAPOMORPHY 28).
It is not possible to determine whether the parallel nearly
vertical ridges along the dorsal surface of the fourth
ceratobranchial, of Curimatopsis are derived, or whether the
divergent, more horizontal processes characteristic of the other
genera of the family are rather apomorphous. The unique nature
of these processes in the Curimatidae among examined
characiforms renders outgroup information useless in polariz-
ing the transition series between the different states. Available
ontogenetic data is similarly uninformative as to which state
is derived. Finally the phyletic distribution of the two character
states within the arrived at phylogeny is such that both the
hypothesis that the Curimatopsis condition is derived and the
alternative hypothesis that the ridge morphology in that genus
is primitive are congruent with the final proposed scheme of
relationships.
The ventral portion of the fourth ceratobranchial also
demonstrates some phylogenetically significant variation
within the Curimatidae. That portion of the bone is unelabo-
rated in most characiform outgroups. In the Curimatidae and
Prochilodontidae, in contrast, there is a distinct process arising
from the medial margin of the ventral surface of the element,
an elaboration which serves as the point of attachment for
various connective tissue bands (Vari, 1983:13) (SYNAPOMOR-
PHY 8). The degree of development of this process varies
drastically within the Curimatidae. Some taxa (Curimatopsis,
Steindachnerina, Pseudocurimata, Curimatella, and
Cyphocharax) have small to very small processes on the ventral
surface of the fourth ceratobranchial, whereas the genera
Potamorhina, Curimata, and Psectrogaster have large, dis-
tinctly anteriorly directed processes on that element. The
distribution of the enlarged processes, the presence of which
is considered derived in light of the smaller processes in the
Prochilodontidae and in other genera of the Curimatidae, is
incongruent with the present phylogeny. Within the context
of the phylogeny proposed herein (see "Synapomorphy List
and Phylogenetic Reconstruction") it is most parsimonious to
consider the acquisition of an enlarged process on the fourth
ceratobranchial a synapomorphy for the clade formed by all
curimatids other than Curimatopsis (SYNAPOMORPHY 57), with
the absence of an enlarged process in Steindachnerina,
Pseudocurimata, Curimatella, and Cyphocharax considered
an apomorphous secondary loss (SYNAPOMORPHY 93). The
alternative and next simplest hypothesis of the independent
gain of an enlarged process in each of Potamorhina, Curimata,
and Psectrogaster involves three ad hoc hypotheses.
LR
LR
MR
LR
A
 B C
FIGURE 26.?Fourth ceratobranchial (C4): A, Curimatopsis microlepis, USNM 268867; B, Psectrogaster ciliata,
USNM 269990; C, Curimata cyprinoides, USNM 267964. (Right side, dorsal view, anterior at top, lateral to
right, dense patterned stippling represents cartilage.)
NUMBER 471 31
Fifth Ceratobranchial (C5)
The fifth ceratobranchial of most characiforms has a lateral
shaft that terminates anteriorly in a cartilage-capped process
lying proximate to the fourth basibranchial. Posteriorly that
shaft ends in a cartilage at the posterolateral corner of the bone.
Along the medial surface of that strut is a flattened plate that
typically has a patch of small conical teeth on its dorsal surface.
Such a form of C5 is common to a variety of members of the
New and Old World Characidae (e.g., Brycon, Weitzman,
1962, fig. 11), the Hemiodontidae (Roberts, 1974, fig. 32),
Lebiasinidae (Weitzman, 1964, fig. 8), Erythrinidae (Roberts,
1969, fig. 35), Hepsetidae (Roberts, 1969, fig. 33), and other
characiforms. As noted by Vari (1983:13), the fifth ceratobran-
chial of the Curimatidae and Prochilodontidae is distinctive in
the significant degree of reduction or total absence of dentition
on the dorsal surface of the ossification (SYNAPOMORPHY 7).
Although prochilodontids lack pharyngeal teeth on the
ventral portions of the gill arches, they nonetheless retain the
overall form of the fifth ceratobranchial generalized for
characiforms that was just described (Figure 27A). All lineages
in the Curimatidae differ from the phylogenetically widespread
type of fifth ceratobranchial found in the Prochilodontidae in
several respects. Members of the Curimatidae have a dorsally
convex medial portion surface of the fifth ceratobranchial
contrary to the flattened, although lateroventrally slanted form
of the bone in examined outgroups (SYNAPOMORPHY 29). An
additional derived character is the overall transverse expansion
of the fifth ceratobranchial. That expansion has two compo-
nents, firstly the widening of the bone medial to the portion
of the bone delimited by the line extending between the anterior
and posterolateral cartilage caps on the element, and secondly
the further development of the ossification lateral to that line
(Figure 27B,C). The transverse expansion of the fifth cera-
tobranchial is an additional synapomorphy for the members
of the Curimatidae (SYNAPOMORPHY 30).
BUCCOPHARYNGEAL COMPLEX
The fleshy lining of the buccopharyngeal region in
characiforms typically consists of a relatively smooth soft
tissue layer conforming to the underlying convex inner surface
of the oral cavity. Such a form of the roof of the mouth is
characteristic of the Prochilodontidae and within the Curimat-
idae is also found in the species of Curimatopsis and
Potamorhina. The genera Curimata, Psectrogaster, Steindach-
nerina, Pseudocurimata, Curimatella, and Cyphocharax have
the soft tissue layers of the buccopharyngeal region elaborated
to various degrees (SYNAPOMORPHY 76). Collectively these
modifications, irrespective of degree of development, are
referred to herein as the buccopharyngeal complex.
The simplest form of the elaboration consists of three
longitudinally oriented folds on the roof of the mouth: a median
fold and a pair of lateral folds. These soft tissue folds begin
somewhat behind the oral valve and terminate distinctly
anterior of the dorsal portions of the gill arches. Such a
relatively simple morphological plan, characteristic of the
majority of curimatids, is dramatically elaborated in two
subunits of the family, the genus Curimata and the majority
of species in Steindachnerina.
All Curimata species have the three primary mouth folds
(MF and paired LF, Figure 28) typical of many curimatids
expanded into large dangling flaps that extend ventrally from
A B C
FIGURE 27.?Fifth ceratobranchial (C5): A, Prochilodus rubrolaenialus, USNM 225419; B, Curimatopsis
microlepis, USNM 268867;, and C, Curimata cyprinoides, USNM 267964. (Right side, dorsal view, anterior at
top, lateral to right, dense patterned stippling represents cartilage.)
32 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
the roof of the oral cavity. These flaps abut against the oral
valve anteriorly and extend posteriorly midway to the anterior
terminus of the dorsal portions of the gill arches (GA). An
additional innovative feature of the buccopharyngeal complex
in Curimata species is the series of secondary folds (SF) on
the roof of the oral cavity. These secondary folds either arise
from the margins of the primary folds, or if independent of
those major flaps, are aligned parallel to them. The greater
degree of development of the median and lateral folds and the
occurrence of secondary folds in the Curimata buccopharyn-
geal complex are considered derived in light of their unique
nature within examined characiforms and synapomorphous for
the genus (SYNAPOMORPHY 82).
The second significant elaboration of the soft tissues of the
roof of the mouth within the Curimatidae involves the lobulate
form of the buccopharyngeal complex characteristic of most
species in Steindachnerina. All nominal members of the genus,
except S. bimaculata, S. semiornata, S. trachysteta, S.
bimaculata sialis, S. pterostrigma, S. melanira, S. conspersa,
S. leucisca leucisca, S. leucisca bolivae, S. argentea, and S.
binotata, have numerous lobulate protruberances extending
ventrally from the roof of the mouth (ALP, Figure 29). These
processes, which totally fill the entire anterior portion of the
roof of the oral cavity, undergo an ontogenetic elaboration from
relatively simple, discrete finger-like projections, to well
developed, thickly packed lobulate structures. The Steindach-
nerina species with such complex protruberances also have in
common elaborate lobed pads (PLP) to each side of the rear
of the buccal cavity in the region between the rear of the
anterior lobulate processes and the forward limit of the dorsal
portion of the gill arches (GA, Figure 29). Those structures are
not developed to a comparable degree in species of Curimata
or in curimatids with less elaborate forms of the anterior portion
of the buccopharyngeal complex.
UL
FIGURE 28.?Roof of the buccopharyngeal chamber and anterior portion of
gill arches of Curimata cyprinoides, USNM 267963, ventral view showing
buccopharyngeal complex (hyoid apparatus, ventral portion of gill arches, eyes,
and associated tissues removed).
FIGURE 29.?Roof of the buccopharyngeal chamber and anterior portion of
gill arches of Steindachnerina hypostoma, USNM 261493, ventral view
showing buccopharyngeal complex (hyoid apparatus, ventral portion of gill
arches, eyes, and associated tissues removed).
NUMBER 471 33
The dangling lobulate processes in most Steindachnerina
species differ significantly in position and detail from the
longitudinal folds characteristic of Curimata. The hypothesis
of homoplasy predicated on the basis of these morphological
differences is also congruent with the most parsimonious
hypothesis of relationships within the Curimatidae. That
phylogenetic scheme indicates that Curimata and the subunit
of Steindachnerina with buccopharyngeal complex elabora-
tions do not constitute a monophyletic lineage, and that the
increased development of the complex in the two lineages was
consequently achieved independently.
The various types of buccopharyngeal complex, be they the
simple three folds of many curimatid species or the more
developed modifications just described for Curimata and most
Steindachnerina species, all serve to increase the surface area
of the dorsal surface of the buccal cavity. The buccal region
of teleosts is well endowed with mucus-secreting cells (see
Kapoor and Evans, 1975) and histological examinations of the
foldings and lobulate bodies in curimatids with well developed
buccopharyngeal complexes shows that the surfaces of these
processes are profusely endowed with mucus-secreting cells
(MPL, Figures 30, 31). Thus the majority of curimatids, in
particular the species of Curimata and most species of
Steindachnerina, presumably have a pronounced increase in
their ability to generate mucus in the buccal cavity, an
adaptation perhaps correlated with their specialized micro-
phagous and detritivorous diet.
The more elaborate forms of the buccopharyngeal complex
in curimatids are similar to the vomero-palatine organs of
some groups of Old World cyprinids recently discussed by
Reid (1982). Although Old World cyprinids and curimatids are
not closely related, they do have in common a microphagous
feeding habit. Reid, refining and further developing a concept
suggested by Matthes (1963), hypothesized that the vomero-
palatine organ in Old World cyprinids functioned in bolus-
formation. He proposed that the buccopharyngeal modifica-
tions served in the production of relatively large amounts of
mucus, and in the mixing of such precipitating mucus with
ingested particles, particularly aufwuchs (a conglomerate of
algae, minute animals, detritus and inorganic matter). Field
observations have shown that curimatids pick at the aufwuchs
which coats many subaquatic surfaces in Neotropical freshwa-
ters, both running and still (pers. observ.). The utilization of
that food source by at least some curimatids is also indicated
by the food item studies of various authors (Azevedo et al.,
1938; de Godoy, 1975; Nomura and Hayashi, 1980; Nomura
and Taveira, 1979). Thus Reid's hypothesis of a particle-
precipitating mucus system in some Old World Ostariophysan
groups is apparently also applicable to the Curimatidae,
paralleling the system used by some microphytophagous
cichlid fishes (Moriarty et al., 1973) and some anuran tadpoles
(Wasserzug, 1972).
HYOID ARCH
Urohyal
The four branchiostegal rays of each side of the head in
conjunction with their associated musculature and branchios-
tegal membranes are an interconnected complex that forms the
ventral limits of the buccal cavity. In many, if not most,
characiforms the branchiostegal rays and associated soft parts
of each side of the head form a flexible complex that is separate
(sensu McAllister, 1968) from its counterpart of the opposite
side. The degree of separation of the complexes of each side
varies among taxa in the order, but in many groups, particularly
those with predatory or omnivorous habits, the opening extends
MPL
FIGURE 30.?Histological section through flaps of buccopharyngeal complex
of Curimata cyprinoides, USNM 267963.
FIGURE 31.?Histological section through lobes of buccopharyngeal complex
of Steindachnerina hypostoma, USNM 261513.
34 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
far anteriorly resulting in an elongate gill aperture on each side.
The condition in the Curimatidae differs from that generalized
bauplan in having a broad connection of the branchiostegal
rays of each side across the ventral midline to the contralateral
complex, and with an attachment of the rays medially to the
urohyal. These interconnections result in a dramatic decrease
in the relative size of the aperture of the gill openings on each
side of the head, and in the degree of mobility in the entire
branchiostegal complex.
A broad contact between the branchiostegal ray complexes
of each side comparable to that in the Curimatidae is also found
in the families Prochilodontidae, Anostomidae, and Chilodon-
tidae, which are the closest relatives of the Curimatidae. Such
an extensive attachment of the complex also occurs in some
characiform outgroups including the Neotropical Parodontidae
and subunits of the Old World family Distichodontidae. The
alterations of the hyohyoidei abductores muscles associated
with the broad contact of the branchiostegal complexes of each
side of the head to each other and to the urohyal, and the
apomorphous nature of those modifications in the Curimatidae,
Prochilodontidae, Anostomidae, and Chilodontidae were dis-
cussed by Vari (1983:45-46, fig. 40). Within that set of four
families the form of the interconnection between the branchios-
tegals and urohyal in the species of the Curimatidae is unique
in the presence of a distinctive highly developed midventral
ligament (Figure 32). That ligament (MVL) is enclosed within
the connective tissue complex joining the medial branchios-
tegal rays (BR) of the two sides and consists of a central
cord-like section with anterior and posterior medially divided
segments. The anterior branch of the ligament of each side is
rounded in cross-section and arises from a relatively broad area
on the ventral surface of the ventral hypohyal (VH). Posteriorly
the central cord-like ligament subdivides into lateral band-like
portions that extend along the posterior margins of the
branchiostegal rays and progressively becomes thinner posteri-
orly. Although the medial ligamentous complex is also found
in the families Chilodontidae, Anostomidae, and Prochilodon-
tidae, it is much less well defined and not as well developed
as the condition in the Curimatidae. Thus the condition in the
last family is considered derived.
The central median cord-like portion of the described
ligamentous complex extends along a distinct groove located
along the ventromedial surface of the urohyal. That urohyal
groove is delimited by ventrolaterally aligned flanges extend-
ing from the lower margin of the main body of that ossification.
The Prochilodontidae, the sister group to the Curimatidae, and
the Anostomidae and Chilodontidae, which together form the
sister group to the curimatid-prochilodontid clade, both have
the ventral portions of the urohyal expanded laterally to
differing degrees; however, in all these families the bone is
ventrally flattened, rather than having a central groove.
Furthermore the lateral flanges in those proximate outgroups
are not ventrolaterally oriented. Thus neither the highly
developed ligamentous system along the midline of the
branchiostegal complexes, nor the groove on the ventral
surface of the urohyal are known to occur in characiforms
outside of the Curimatidae. The complex formed by the
midventral ligament and the urohyal is consequently hypothe-
sized to be a synapomorphy for the members of the
Curimatidae (SYNAPOMORPHY 31).
The form of the urohyal is quite variable within the
Curimatidae, both in terms of overall morphology and in
various details. The continuum between most different
morphological states makes it difficult, if not impossible, to
non-arbitrarily delimit much of that variability in terms of
discrete characters applicable to a phylogenetic analysis.
Several distinct character states exist which are, however,
useful in generating a hypothesis of generic level relationships.
As noted above, the urohyal of the Curimatidae is
characterized by a posteroventrally angled flange along both
sides of its ventral margin. In cross section the bone
consequently has the form of an inverted "Y"; the ventral arms
of which vary in the degree of lateral development within the
family. In Curimatopsis the ventrolateral flanges are quite
narrow along their entire length, whereas in the other lineages
AC
MVL
BR
FIGURE 7,2,?Cyphocharax abramoides, USNM 267953, hyoid arch, anterior
portion of branchiostegal rays and midventral ligament (medial portion and
right side, ventral view, anterior at top, basihyal not illustrated, halves of arch
separated laterally from normal positions).
NUMBER 471 35
of the Curimatidae, the flanges are wider, sometimes
considerably so, particularly anteriorly. Outgroup comparisons
reveal wide flanges on the urohyal in the Prochilodontidae, the
sister group to the Curimatidae. The sister clade to the lineage
formed by those two families shows two divergent conditions;
narrow flanges in the Anostomidae and wide processes in the
Chilodontidae. The condition in the immediate outgroup, the
Prochilodontidae, together with parsimony considerations in
the broader outgroup, including the Anostomidae and Chi-
lodontidae, support the hypothesis that a urohyal with wide
ventrolateral flanges is primitive for the family Curimatidae.
Thus the narrow processes in the species of Curimatopsis is
considered a reductive feature that is synapomorphous for the
members of that genus (SYNAPOMORPHY 45).
The urohyal in the Prochilodontidae, Anostomidae, and
Chilodontidae are all relatively short longitudinally with their
ventral portions terminating approximately at the point where
the medialmost branchiostegal rays diverge laterally. The
Curimatidae, in contrast, has the ventral portions of the urohyal
extending moderately to distinctly beyond that point, a
condition that is thus considered a synapomorphy for the
members of the family (SYNAPOMORPHY 32).
Within the Curimatidae, the genera Curimatopsis, Pota-
morhina, and Psectrogaster all have a urohyal that is distinctly
elongate, with one-third to one-half of the ossification
extending beyond the rear of the point of the lateral divergence
of the medial branchiostegal rays. Such a degree of posterior
development of the urohyal is not found in the remainder of
the family, and is considered derived on the basis of outgroup
comparisons. The possession of an elongate urohyal does not,
however, delimit a monophyletic subunit of the family in the
final proposed intrafamilial phylogeny. Within that phyletic
scheme (Figure 44), Curimatopsis and Potamorhina are
sequential clades but are separated from Psectrogaster by the
lineage consisting of Curimata. The sister group to Psectrogas-
ter, the assemblage consisting of Steindachnerina, Pseudocuri-
mata, Curimatella, and Cyphocharax is characterized by short
urohyals. Three equally parsimonious hypotheses exist to
explain the distribution of the derived condition, the elongate
urohyal, within the Curimatidae. First, there may have been
independent acquisition of the modification in the three
lineages, Curimatopsis, Potamorhina, and Psectrogaster.
Second, there may have been a single synapomorphic
acquisition at the base of the Curimatidae, with secondary
shortening of the urohyal occurring independently in Curimata
on the one hand and the assemblage consisting of Steindach-
nerina, Pseudocurimata, Curimatella, and Cyphocharax on the
other. Finally, the elongation may have been acquired at the
level of Curimatidae, with a secondary shortening occurring
in the clade consisting of Curimata, Psectrogaster, Steindach-
nerina, Pseudocurimata, Curimatella, and Cyphocharax, with
a secondary lengthening in Psectrogaster.
Each of these hypotheses involves three ad hoc hypotheses
about evolutionary events. Given these equally parsimonious
alternatives, it was not possible to determine the appropriate
phylogenetic levels at which to use these characters as
synapomorphies; consequently, they are not incorporated in
the cladogram (Figure 44), although the feature was included
in the PAUP analysis and has been assigned a synapomorphy
number (SYNAPOMORPHY 105) for analysis and discussion
purposes.
Basihyal (BH) and Basihyal Tooth Plate (BHTP)
The basihyal is an unpaired median element that extends
anteriorly from between the paired dorsal hypohyals. The
element typically consists of a cartilaginous core, at least in
juveniles, enveloped posteriorly and to varying degrees
anteriorly by a sheath of bone. Along its dorsal surface the
basihyal may or may not be partially covered by an
independent ossification, the thin basihyal tooth plate, that
extends anteriorly over the dorsal surface of the cartilaginous
portion of the element (Figure 33B). Although the form of the
complex formed by the basihyal and basihyal tooth plate varies
considerably within the family, the continuum between the
majority of morphological states permits the delimitation of
only one evidently derived character state.
The majority of curimatids have an elongate to elongate
triangular basihyal complex with relatively straight lateral
margins (Figure 33A). In Steindachnerina species, rather, the
bone-cartilage complex is apomorphously laterally expanded
anteriorly, with the anterior cartilaginous portion notably flared
outwards. Associated with this reconfiguration of the basihyal
is a comparable lateral expansion of the basihyal tooth plate
(BHTP, Figure 33B). The anterior expansion of the basihyal
and basihyal tooth plate are apomorphous for the species
Steindachnerina (SYNAPOMORPHY 100). A similarly expanded
basihyal-basihyal tooth plate complex also occurs in the
BHC BHC
BHTP
B
FIGURE 33.?A, Basihyal erf Potamorhina altamazonica, USNM 257367; B,
basihyal and basihyal tooth-plate of Steindachnerina bimaculata, USNM
261450. (Dorsal view, anterior at top, dense patterned stippling represents
cartilage.)
36 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
Prochilodontidae, the sister group to the Curimatidae. In light
of the unique nature of the Steindachnerina form of the
basihyal complex within the Curimatidae and the topology of
the most parsimonious phylogenetic scheme in that family
(Figure 44), the hypothesis that the expansion of the basihyal
and its associated tooth plate in the Prochilodontidae and
Steindachnerina represent homoplasies is the most parsimoni-
ous.
A basihyal tooth plate is present in the Anostomidae (see
Winterbottom, 1980, fig. 45), Prochilodontidae (see Van,
1983, fig. 22) and Chilodontidae (see Vari, 1983, fig. 24), the
immediate sister groups to the Curimatidae. Within the
Curimatidae the ossification is variously present at the generic
and subgeneric level. Although lacking in Curimatopsis,
Potamorhina, and Psectrogaster, that element is found in a
minority of Curimata species and in the large majority of the
species of Steindachnerina, Pseudocurimata, Curimatella, and
Cyphocharax. It would thus appear that the primitive condition
within the family is for the ossification to be absent. The
presence of a basihyal tooth plate in most members of
Steindachnerina, Pseudocurimata, Curimatella, and Cyphocharax
is thus considered to be indicative of the monophyly of that
assemblage, (SYNAPOMORPHY 94) albeit with qualifications
due to the absence of the tooth plate in some members of that
lineage.
modifications of the branchiostegal rays that differ from the
morphological plan just described. In the Curimatidae the
anterodorsal portion of the first and second branchiostegals
(BRj and BR2) is dorsally expanded resulting in a pronounced
angle in the anterior margin of those bones. As a consequence
the margins of these ossifications are distinctly concave
proximally, more so in the case of the first branchiostegal ray.
The anterodorsal expansion of the proximate portion of the
third branchiostegal (BR3) of curimatids is even more
pronounced. This expansion results in the third branchiostegal
ray distinctly overlapping the dorsal portion of BR2 laterally
and completely covering the section of that element articulating
with the anterior ceratohyal (Figure 34B). In light of the
conditions in the proximate outgroups, these modifications of
the portions of the first to third branchiostegal rays approximat-
ing the hyoid arch are considered apomorphous for the
members of the Curimatidae (SYNAPOMORPHY 33).
ANTORBITAL AND INFRAORBITALS
The orbit in the vast majority of characiforms is ringed by
a series of eight bones, an antorbital, a supraorbital, and six
infraorbitals, with the infraorbitals bearing portions of the
laterosensory canal system. Variations in the presence or
Branchiostegal Rays (BRj_4)
The Curimatidae, Prochilodontidae, Chilodontidae,and Anos-
tomidae all typically have four branchiostegal rays, although
that number is reduced to three in some anostomids
(Winterbottom, 1980). The condition in the prochilodontid
Semaprochilodus laticeps (Figure 34A) is representative of the
generalized condition of the portions of those elements
proximate to the hyoid arch in that family. That morphological
plan also represents the plesiomorphous condition of those
portions of the branchiostegal rays for the Chilodontidae and
Anostomidae with respect to the characters of interest in this
discussion.
The first two branchiostegal rays in the Prochilodontidae
(BR, and BRj) attach to a depression on the ventral surface of
the anterior ceratohyal. According to the degree of ossification
of the entire hyoid arch, a largely size related function, the area
of attachment either has the form of two ossified notches
(Figure 34A) or is a broader concave region subdivided by
smaller cartilages (Figure 34B). Anterodorsally the margins of
the first two branchiostegal rays (BRj and BR2) are slightly
convex or straight, but without any pronounced angle
proximate to the point of articulation. The third branchiostegal
ray (BR3) attaches to the posterolateral surface of the anterior
ceratohyal and only slightly overlaps the posterolateral margin
of the second branchiostegal ray leaving the portion of that
bone proximate to the anterior ceratohyal exposed laterally.
All members of the Curimatidae demonstrate several derived
PC
BR
BR2 BR3
FIGURE 34.?Hyoid arch (without interhyal) and proximate portion of
branchiostegal rays: A, Semaprochilodus laticeps, USNM 270239; and B,
Steindachnerina bimaculata, USNM 261450. (Left side, lateral view, anterior
to left, dense patterned stippling represents cartilage.)
NUMBER 471 37
absence of some of these elements, their relative sizes, and
presence and form of the included laterosensory canal segments
serve to define subunits of the Curimatidae.
Antorbital
The antorbital (the adnasal of Gregory, 1933, and Daget,
1964), an element lying immediately anterior to the ventral
process of the lateral ethmoid, demonstrates only one
phylogenetically significant character in the Curimatidae. As
noted by Vari (1982a:8) the antorbital is absent in all species
of Curimatopsis. The possibility that the absence of the
antorbital in Curimatopsis is the result of some paedomorphic
process would in first consideration appear likely in light of
the relatively small adult size of most species of Curimatopsis.
Such an explanation also seems applicable in the case of the
diminutive African characid Lepidarchus adonis Roberts
which also lacks an antorbital (Roberts, 1966). In Curimatop-
sis, however, the ossification is absent not only in the
diminutive members of the genus such as C. evelynae, but
even in a 75.8 mm SL specimen of C. microlepis (USNM
268867) cleared and counterstained for cartilage and bone.
Thus the absence of the bone in all Curimatopsis species does
not appear to simply represent a paedomorphic feature.
Weitzman and Vari (1988) noted that paedomorphosis must
be analyzed within a phylogenetic context in which miniaturi-
zation and miniatures are evaluated relative to the size and
features of those taxa phylogenetically proximate to the
potential miniatures. Relative to the much larger body sizes
of the Prochilodontidae, and the immediate sequential sister
groups to Curimatopsis in the Curimatidae, even the largest
Curimatopsis species is of a reduced overall size. Regardless
of the underlying cause, the lack of an antorbital is considered
a synapomorphy for Curimatopsis species (SYNAPOMORPHY
41).
First Infraorbital (IOj)
The first infraorbital, the most anterior of that series of bones
ringing the ventral and posterior portions of the orbit typically
has a discrete tube-like laterosensory canal segment in
Characiformes. Within the Curimatidae, a laterosensory canal
segment in the first infraorbital is present in the vast majority
of species, but is absent in the five species of Curimatopsis,
four species of Cyphocharax (gillii, saladensis, punctata,
vanderi), and a single species in Steindachnerina (pinotata).
On the basis of the overall phylogeny, the lack of a
laterosensory canal segment in the first infraorbital is
considered a synapomorphy that arose independently in
Curimatopsis on the one hand (SYNAPOMORPHY 46) and the
cited species of Cyphocharax on the other, and which is an
autapomorphy for Steindachnerina binotata.
Gregory and Conrad (1938, fig. 28) in their drawing of an
unspecified "Curimata" species, illustrate a single canal
bearing bone anterior of the lateral ethmoid (their prefrontal
plus paraethmoid). That bone, which they identify as the
lacrymal, is shown as extending from the supraorbital dorsally
around the anterior and anteroventral margins of the orbital rim
to the second infraorbital (their second suborbital) posteriorly.
The illustrated lacrymal of those authors thus appears to
actually represent a combination of two separate bones, the
antorbital (that portion of their ossification proximate to the
lateral ethmoid) and the first infraorbital (the ventral canal-
bearing section of their illustration). Their illustration is also
incorrect in showing the infraorbital series posterior of the first
infraorbital as a single continuous ossification. In all examined
curimatids the second through sixth infraorbitals are discrete
elements.
Fourth and Fifth Infraorbitals (IO4 and IO5)
The fourth infraorbital commonly contains a tripartite canal
in the near relatives of the Curimatidae, the Prochilodontidae
(e.g., Ichthyoelephas Posada-Arango, Roberts, 1973, fig. 17),
Anostomidae (e.g., Synapotalaemus Myers and Fernandez-
Yepez, Winterbottom, 1980, fig. 32) and Chilodontidae (e.g.,
Caenotropus, Gery, 1964b, fig.l). The primary canal segment
in this element extends parallel to the anterior margin of the
bone along the orbital rim and communicates dorsally and
ventrally with comparable canal segments in the fifth and third
infraorbitals respectively. The fourth infraorbital also has a
horizontal or slightly posteroventrally inclined, posteriorly
directed side branch (HC-IO4) of the laterosensory canal
segment. That canal extends posteriorly to the proximity of the
opening in the preopercle into the preopercular laterosensory
canal system, and sometimes communicates posteriorly with
an unossified laterosensory canal segment in the tissue of the
adipose eyelid overlying the opercle. The laterosensory canal
segment of the fifth infraorbital continues along the arch of the
canal segments in the third and fourth infraorbital, without any
pronounced angle between the primary axes of the sensory
canal segments in the fourth and fifth infraorbitals.
The pattern described above is most common in the
Prochilodontidae, Anostomidae, and Chilodontidae, with
differences among those families relative to this pattern
involving features not pertinent to the Curimatidae. Among
curimatids a morphological plan of the fourth and fifth
infraorbitals comparable to that of the cited outgroups occurs
in Curimatopsis, Potamorhina, Curimata, and Psectrogaster
(Figure 35A,B). The genera Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax, in contrast, have the anterior
canal segment of the fourth infraorbital (IO^ more posteriorly
angled, with an associated reduction or elimination of the
posterior horizontal sidebranch in that bone (HC-IO4, Figure
35c). The laterosensory canal segment in the fifth infraorbital
(IO5), in turn, is more distinctly angled posteroventrally than
in the generalized characiform condition. Together, reorienta-
tions of these canals are most obvious in the distinct angle
38 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
IOC
A B C
FIGURE 35.?Fourth and fifth infraorbital and dorsal portion of third infraorbital: A, Psectrogaster curviventris,
USNM 243221; B, Potamorhina altamazonica, USNM 257367; and C, Curimatella meyeri, USNM 261508.
(Left side, lateral view, anterior to left.)
SPH
B D
FIGURE 36.?Sixth infraorbital (dermosphenotic): A, Psectrogaster amazonica, USNM 261518; B, Potamorhina
altamazonica, USNM 257367; c, Curimata cyprinoides, USNM 267964; and D, Steindachnerina conspersa,
USNM 232224. (Left side, lateral view, anterior to left.)
NUMBER 471 39
between the primary axes of the laterosensory canal segments
in the fourth and fifth infraorbitals; a condition which contrasts
with the continuous arch of the canals in those elements in
outgroups within (Figure 35A,B) and outside of the Curimatidae
(SYNAPOMORPHY 95).
Most species of Potamorhina retain the plesiomorphous
orientation of the fourth and fifth infraorbitals and their canals.
The fourth and fifth infraorbitals in that genus are, however,
unique within the Curimatidae in their pronounced posterior
development. This is particularly the case for the fourth
infraorbital and the horizontal laterosensory canal segment in
that ossification, with these shifts in the relative positions of
those canals considered together to be an apomorphy for the
genus (SYNAPOMORPHY 45). Potamorhina pristigaster, which
shares with its congeners the posterior expansion of the fourth
infraorbital also has the form of the laterosensory canal
segments in the fourth and fifth infraorbitals autapomorphously
modified. That species has a distinct acute angle between the
laterosensory canal segments in the fourth and fifth infraorbi-
tals reminiscent of the condition in Steindachnerina, Pseudocu-
rimata, Curimatella, and Cyphocharax. That decreased angle
between the axes of the canals in these bones is a consequence
of the posterior repositioning of the two ossifications relative
to the orbital rim. The condition differs significantly from the
reduced angle between the canal segments in Steindachnerina,
Pseudocurimata, Curimatella, and Cyphocharax which results
from the reorientation of the canal segments in the fourth and
fifth infraorbitals which have not themselves been shifted
posteriorly from the orbital rim in those taxa. Thus the P.
pristigaster condition is considered an autapomorphy non-
homologous with the situation in Steindachnerina, Pseudocuri-
mata, Curimatella, and Cyphocharax.
Sixth Infraorbital (IO^
The sixth infraorbital (the dermosphenotic of some authors)
is an independent ossification that roofs over the dilatator fossa
on the lateral surface of the neurocranium. The generalized
condition of the ossification in characiforms has a tripartite
laterosensory canal system comparable to that of Figure 36A,B,
with the ventral arm of the complex contacting the laterosen-
sory canal segment in the fifth infraorbital (IO5). The
posterodorsal branch of the IO6 canal system (PD) approxi-
mates the anterior opening of the pterotic sensory canal, and
the anterior branch extends anterodorsally to the frontal. Such
a condition of the laterosensory canal system in the sixth
infraorbital is the basic pattern in the Prochilodontidae,
Anostomidae, and Chilodontidae, along with a variety of other
characiforms including members of the Old World families
Citharinidae (Citharinus, Daget, 1962a, fig, 7; Citharidium
Boulenger, Daget, 1962b, fig. 10), Distichodontidae {Xeno-
charax, Daget, 1960, fig. 7; see also Vari, 1979:296-301 for
the remainder of the family), Hepsetidae (Roberts, 1969, fig.
1), and Characidae {Hydrocynus, Allis, 1905, fig. 13). In the
Neotropical realm a similar morphology of the sixth infraorbi-
tal is found in many, although not all, members of the
Characidae (e.g., Brycon, Weitzman, 1962, figs. 8, 9),
Hemiodontidae, and Erythrinidae (Q.g.,Hoplias, Roberts, 1969,
fig. 3) among other families.
Several modifications of the sixth infraorbital and its
enclosed laterosensory canal segment define subunits of the
Curimatidae. The relative size of the sixth infraorbital is
reduced in Curimatopsis microlepis and C. macrolepis. The
three remaining species of the genus (evelynae, crypticus, and
myersi) demonstrate no trace of an ossified sixth infraorbital.
As adults these latter species are some of the smallest in the
family and it is possible that the reduced sixth infraorbital
represents a paedomorphic event characteristic of the clade.
That possibility is particularly notable since the sixth
infraorbital is absent in a number of diminutive characiforms
(see Weitzman and Fink, 1983, fig. 24, and Weitzman and
Fink, 1985:63, fig. 54). Nonetheless it is noteworthy that
cleared and stained specimens of Curimatopsis crypticus of
42.1 mm SL show no indication of an ossified sixth infraorbital
although the element is present in comparably sized individuals
of other curimatid species that achieve larger adult body sizes.
Regardless of the process underlying the lack of an ossified
dermosphenotic in these taxa, that absence is considered
derived and a synapomorphy for the clade consisting of
Curimatopsis evelynae, C. myersi, and C. crypticus {Curima-
topsis SYNAPOMORPHY F). Similarly the combination of a
reduction or loss of that bone is hypothesized to be
synapomorphous for the entire genus (SYNAPOMORPHY 42).
Within the Curimatidae, two clades at the generic and
suprageneric levels are definable by derived states of the sixth
infraorbital and its contained laterosensory canal segment. The
elaboration of the plesiomorphous tripartite IO6 laterosensory
canal segment via the addition of a posteroventral and
sometimes also anteriorly and anteroventrally aligned side
branches (Figure 36c) was discussed by Vari (1984a:6). That
condition, synapomorphous for the species of Potamorhina
(SYNAPOMORPHY 67), is not analyzed again here.
Whereas a tripartite canal system in the sixth infraorbital or
an elaboration of that pattern are common to species of
Curimatopsis with that element, along with Potamorhina,
Curimata, and Psectrogaster, all other curimatids have only a
single unbranched laterosensory canal segment in that ossifica-
tion (Figure 36D). That simplified straight canal segment
communicates with the fifth infraorbital sensory canal (IO5)
ventrally, and extends dorsally to approximate the anterior
opening of the pterotic sensory canal, lacking any contact with
the laterosensory canal segment in the frontal. Given the
generality of a tripartite laterosensory canal segment in the
sixth infraorbital of proximate curimatid outgroups and among
characiforms as a whole, the simple laterosensory system in
the sixth infraorbital in Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax is hypothesized as derived
(SYNAPOMORPHY 96).
40 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
OPERCULAR APPARATUS
Opercle and Subopercle
The close morphological and functional association of the
opercle and subopercle makes it most efficient to discuss these
elements jointly. The Curimatidae as a whole demonstrates a
posteroventral truncation of the opercle and the associated
expansion of the subopercle relative to the condition of these
elements in examined outgroups. These modifications are,
however, difficult to quantify non-arbitrarily, and thus I prefer
not to utilize them as synapomorphies for the members of the
Curimatidae, a clade that is already delimited by numerous
uniquely derived characters. Within the Curimatidae, the
opercle is significantly expanded posteriorly in all Pota-
morhina species relative to the condition in the outgroups.
That expansion of the opercle in conjunction with the overall
posterodorsal reorientation of the originally vertical primary
axis of the opercle results in a pronounced inclination of the
line of junction between the opercle and subopercle (see Van,
1983, fig. 22). These modifications in the overall form and
relative position of the opercle are together considered to be a
synapomorphy for the species of Potamorhina (SYNAPOMOR-
PHY 73).
Interopercle
The interopercle, the element lying ventral of the preopercle,
and anterior of the subopercle and the anteroventral portion of
the opercle is modified to some degree in all curimatids relative
to the condition in the outgroups and also demonstrate a degree
of variability within the family. In proximate outgroups to the
Curimatidae the ossification is only vertically developed to a
moderate extent in the region where it abuts the opercle and
subopercle (see Winterbottom, 1980, fig. 33 for Anostomidae;
Gery, 1964b, fig. 1 for Chilodontidae; and Van, 1983, fig. 28
for Prochilodontidae). A somewhat more pronounced develop-
ment posteriorly of the interopercle is typical for all curimatids.
The intrafamilial variation in the form of the interopercle makes
it difficult to non-arbitrarily define discrete characters within
that continuum. Thus this feature is not used as a synapomor-
phy for the species of the family, which share numerous other
synapomorphies.
The continuum between the forms of the interopercle within
the Curimatidae also makes it difficult to non-arbitrarily
discriminate character states for this element. As a consequence
it is possible to only recognize readily a single derived
condition of the interopercle within the family. In all
Potamorhina species the interopercle is significantly further
developed posteriorly into a triangular plate filling the space
between the rear of the preopercle and the anterior margin of
the opercle and subopercle (see Vari, 1983, fig. 22). This
apomorphous expansion is considered a synapomorphy for
Potamorhina (SYNAPOMORPHY 74).
LlGAMENTUM PRIMORDIALE
The Aj section of the adductor mandibulae muscle in
curimatids extends anteriorly from the preopercle to an
attachment on the posterdorsal portion of the lower jaw.
Arising from the anterodorsal portion of the adductor
mandibulae and the proximate portions of the angulo-articular
is a tendonous band that extends anterodorsally to an insertion
either on the posterior margin {Curimatopsis species) or the
medial surface of the main portion of the maxilla (all other
curimatids). The ligament gradually tapers dorsally, without
any marked shift in form or size. That band, the ligamentum
primordiale (the primordial ligament of Winterbottom, 1974:232,
and the articular-maxillary ligament of Alexander, 1964:183),
arises in most characiforms from an aponeurotic sheet on the
anterodorsal portion of Aj and the proximate portions of the
angulo-articular.
Although the area of origin of the ligament and the degree
of its association with the angulo-articular varies within
characiforms, the described morphology of the band is typical
for all examined outgroups in the order. Within the Curimat-
idae such a relatively straight, cord-like ligamentum primor-
diale is found only in the species of Curimatopsis. All the
remaining members of the family have the ligamentum
primordiale complex modified to varying degrees. The species
of Potamorhina and Curimata have the aponeurotic sheet on
the anterodorsal portion of Ay that typically serves as the origin
of the ligamentum primordiale greatly expanded into a discrete
thick connective tissue mass at the base of the main portion
of the ligament. That thickening lies proximate to and is
anteroventrally attached by connective tissue to the posterodor-
sal margin of the angulo-articular but structurally remains
relatively distinct from the lower jaw.
This expanded ventral portion of the ligamentum primor-
diale complex in Curimata and Potamorhina is further
modified in the genera Psectrogaster, Steindachnerina, Pseu-
docurimata, Curimatella, and Cyphocharax. Those taxa have
a discrete dense mass centered along the longitudinal axis of
the ligament. That body largely replaces the expanded fibrous
connective tissue body found on the proximal section of the
ligamentum primordiale in Curimata and Psectrogaster.
Histological examination has shown that the body included in
the ligamentum primordiale consists of cartilage enclosed
within a aponeurotic sheet of varying thickness (Figure 37).
The anterior portion of the cartilage serves as an area of
attachment for a relatively extensive portion of the ligamentum
primordiale. Posteriorly the cartilage body is continuous with
the tendenous fibers of the anterior portion of the Aj section
of the adductor mandibulae muscle. The entire cartilage-
ligament complex attaches via loose ligamentous bands to the
posterodorsal margin of the angulo-articular. As is the case of
the thickened connective tissue body in Curimata and
Potamorhina, the ligamentum primordiale cartilage in Psec-
trogaster, Steindachnerina, Pseudocurimata, Curimatella, and
NUMBER 471 41
CB-LP
DEN
FIGURE 37.?Ligamentum primordiale, cartilage body of ligamentum primor-
diale, and proximate portions of Aj section of adductor mandibulae muscle,
maxilla, dentary, and angulo-articular of Steindachnerina bimaculata, USNM
261450. (Jaws in open position, continuation of ligamentum primordiale
medial to maxilla indicated by dashed lines, right side, lateral view, anterior
to right, dense patterned stippling represents cartilage.)
Cyphocharax is not directly associated with the lower jaw.
These modifications and elaborations of the proximal
portions of the ligamentum primordiale are unique to the cited
genera of the Curimatidae among examined characiforms and
are considered to represent synapomorphies at two different
levels of universality within the Curimatidae. The possession
of some form of elaboration of the proximal portion of the
ligamentum primordiale, either with or without an included
cartilage body, is hypothesized to be a synapomorphy for the
genera Potamorhina, Curimata, Psectrogaster, Steindachner-
ina, Pseudocurimata, Curimatella, and Cyphocharax (SYNAPO-
MORPHY 58). The presence of a cartilage mass in the
ligamentum primordiale connective tissue complex is, in turn,
hypothesized to be a further derived synapomorphy for the
assemblage consisting of Psectrogaster, Steindachnerina,
Pseudocurimata, Curimatella, and Cyphocharax (SYNAPO-
MORPHY 84).
UPPER AND LOWER JAWS
The most obvious distinguishing character of the curimatid
jaws, upper and lower, is the absence of oral dentition in adults
of the family. Juveniles of Curimata cyprinoides of up to
approximately 25 mm SL have a single row of unicuspidate
teeth in each jaw. Comparable dentition has been reported for
Steindachnerina elegans by Azevedo et al. (1938, figs. 1, 2)
and for several upper Rio Parana" curimatid species by de
Godoy (1975:589, 596, 601). As juveniles of the Curimatidae
alter their diet from zooplankton to detritus, this larval dentition
is completely lost in all specimens above 30 mm SL. The vast
majority of characiforms have some form of dentition, typically
well developed, as adults and the absence of jaw teeth in adult
curimatids is considered a synapomorphous reduction for the
family (SYNAPOMORPHY 16).
The only other members of the Characiformes lacking jaw
teeth as adults are Anodus elongatus Spix and Eigenmannina
melanopogon (Cuvier) of the Hemiodontidae. As noted under
"Synapomorphy List and Phylogenetic Reconstruction," this
similarity is considered to represent a homoplasious loss of
teeth in the hypothesized ancestoral species of the Curimatidae
and in the lineage consisting of Anodus and Eigenmannina in
the Hemiodontidae rather than a synapomorphy of Curimatidae
and those taxa.
Premaxilla (PMX)
The premaxillae of curimatids are joined across the midline
by connective tissue bands and attached to the ethmoid both
through broad connective tissue sheets and via discrete
ligaments. The range in the relative size and overall position
of the mouth in different subunits of the family is reflected in
the variety of premaxillary forms found among curimatids.
Most of this variation, however, forms a continuum that is
difficult, if not impossible, to non-arbitrarily subdivide into
discrete units and thus it is not readily used in phylogenetic
analyses. Two different discrete modifications of the premax-
illa are recognizable, however, and serve to delimit assem-
blages of genera within the Curimatidae.
The premaxilla of the proximate outgroups to the Curimat-
idae, the families Prochilodontidae, Anostomidae and Chi-
lodontidae, are all apomorphously modified in overall form
(Vari, 1983:9) making them inappropriate as outgroups for
evaluating the plesiomorphous form of the premaxilla among
curimatids. As a consequence more distant outgroups, albeit
of inexact phyletic position relative to the Curimatidae, were
utilized. These examinations have shown that the generalized
condition of the premaxilla in those characiform outgroups is
a relatively well developed bone with a definite triangular form
in dorsal or anterodorsal view. Such a premaxilla is
characteristic of the vast majority of curimatids. The members
of the genus Curimatopsis, alternatively, have an extremely
elongate premaxilla that gradually tapers laterally to a curved
rod-like shaft. The unique nature of this form of premaxilla
within the Curimatidae and proximate outgroups results in its
being considered a synapomorphy for the members of
Curimatopsis (SYNAPOMORPHY 47).
The second alteration of the premaxilla that characterizes a
subunit of the Curimatidae involves the mode of articulation
between the premaxilla and maxilla. The posteroventral margin
of the premaxilla extends a short distance lateral to the anterior
margin of the main body of the maxilla. The point of contact
between those elements in Potamorhina, Psecterogaster,
Steindachnerina, Pseudocurimata, Curimatella, and Cyphocharax
42 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
is elaborated into a discrete joint with an articular facet on both
the premaxilla and maxilla at their point of contact. These
modifications of the two bones at their area of articulation is
considered to be a synapomorphy for the genera Potamorhina,
Curimata, Psectrogaster, Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax given the absence of such
structures in the examined characiform outgroups (SYNAPOMOR-
PHY 59).
Maxilla (MX)
The maxilla, the second element of the upper jaw,
demonstrates a number of discrete modifications within the
Curimatidae, several of which are pertinent to the generic-level
phytogeny in the family. Outgroup comparisons to determine
the plesiomorphous condition of the maxilla within the
Curimatidae are obscured by the apomorphous modifications
of the maxilla in each of the proximate sister groups of the
Curimatidae (Prochilodontidae, Anostomidae, and Chilodon-
tidae). Outgroup comparisons to more distantly related
characiform taxa of uncertain phylogenetic relationship relative
to the clade formed by those three families and the Curimatidae
nonetheless allow the advancement of a hypothesis of the
generalized condition of the bone. Based on those comparisons,
the generalized characiform condition of the maxilla is
considered herein to be a moderately developed, laterally
flattened blade-like body ventrally which is continuous
dorsally with the dorsal rod-like articular process (e.g., various
glandulocaudine species, Weitzman and Fink, 1985, figs.
59-66). Such a form of the maxilla is not encountered in the
Curimatidae but a variety of derived overall morphologies of
the bone do characterize subunits of the family.
Curimatopsis has a paddle-shaped, laterally flattened main
body of the maxilla. The attachment of the ligamentum
primordiale (LP) onto the maxilla is at the rear of the
ossification in the area where the dorsal process of the maxilla
joins the main body of the bone (AI, Figure 38A). Such an
area of attachment is generalized for examined characiforms
including the Anostomidae and Chilodontidae and is thus
considered plesiomorphous. As a consequence of the highly
modified form of the oral jaws in the Prochilodontidae, it is
not possible to utilize the condition in that family for outgroup
comparisons. The anteriorly expanded paddle-shaped main
body of the bone characteristic of Curimatopsis, in contrast,
is encountered in only a few characiform outgroups (e.g.,
Thrissobrycon pectinifer Bohlke, 1953, fig. 1). In light of its
unusual nature within characiforms the condition of the maxilla
in Curimatopsis is consequently considered a synapomorphy
for the species of Curimatopsis (SYNAPOMORPHY 48).
The species of Potamorhina, Curimata, Psectrogaster,
Steindachnerina, Pseudocurimata, Curimatella, and Cyphocharax
have a dramatically different morphology of the maxilla. In all
the species of those genera the anterior margin of the bone is
only slightly developed, often with a notch to accommodate
FIGURE 38.?Maxilla and anterodorsal portion of ligamentum primordiale: A,
Curimatopsis microlepis, USNM 268867; and B, Psectrogaster ciliata, USNM
269990. (Left side, lateral view, anterior to left, area of insertion of ligamentum
primordiale on maxilla not obvious in lateral view, point on lateral surface
overlying insertion indicated by "AF'.)
the lateral portion of the premaxilla (N, Figure 38B). The
posterodorsal portion of the main body of the maxilla,
alternatively, is developed into a large plate-like process (PP)
that extends under the first infraorbital when the mouth is
closed. The attachment of the ligamentum primordiale in these
taxa (AI) is now on the medial surface of the maxilla at the
anterior margin of this expanded posterodorsal process.
Utilizing that point of attachment as a homologous landmark,
it appears that the posterodorsal process of the maxilla is the
consequence of an expansion of that region of the bone. Given
the absence of such a process on the maxilla in other examined
characiforms, the possession of that structure is considered
derived and a synapomorphy for the genera Potamorhina,
Curimata, Psectrogaster, Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax (SYNAPOMORPHY 60). As noted
above, these taxa also have in common the derived possession
of an articular facet on the dorsal margin of the maxilla which
contacts a comparable process on the premaxilla.
Dentary (DEN)
The dorsal margin of the dentary bone typically is anteriorly
straight or slightly convex, and posteriorly distinctly convex
in the proximate outgroups to the Curimatidae, the families
Prochilodontidae, Anostomidae and Chilodontidae (see Vari,
1983, fig. 2). That condition, hypothesized as plesiomorphous
for the Curimatidae, is also found in the family in all members
of Curimatopsis (Figure 39A). The remainder of the family
Curimatidae, in contrast, has a sigmoid dorsal margin to the
dentary, with the convex central portion of the bone flanked
by two concave regions (Figure 39B). The centrally convex
region, hypothesized as derived relative to the condition in
Curimatopsis and the outgroups to the Curimatidae, is thus
NUMBER 471 43
DEN
B
DE1 AA
FIGURE 39.?Lower jaw: A, Curimatopsis microlepis, USNM 268867; and B,
Curimatella alburna, MZUSP 6309. (Left side, lateral view, anterior to left.)
considered a synapomorphy for Potamorhina, Curimata,
Psectrogaster, Steindachnerina, Pseudocurimata, Curimatella,
and Cyphocharax (SYNAPOMORPHY 61). It demonstrates
notable variation within the family. In some taxa (e.g.,
Curimata) it is developed into a distinct, somewhat triangular
process, whereas in others (e.g., Potamorhina) it curves
laterally into a bony lip. The non-Curimatopsis. curimatids
cannot feasibly be non-arbitrarily divided into discrete subsets
based on this variation. I consequently will not utilize those
differences at less inclusive phylogenetic levels.
PALATINE ARCH
The palatine arch consists of four bones, the palatine,
ectopterygoid, mesopterygoid, and metapterygoid, and their
associated cartilages. In the majority of characiforms those
elements form a closely interconnected complex extending
between the upper jaw, ethmo-palatine cartilage and vomer
anterodorsally, the quadrate ventrally and the hyomandibula
posteriorly. Each of the components of the palatine arch
demonstrates derived modifications pertinent to the question
of the monophyly of the Curimatidae and various subunits of
that family. In order to simplify the discussion and evaluation
of the polarities of each of these alterations, the morphology
of each component of the arch will be discussed individually.
Palatine (PAL) and Ethmo-palatine Cartilage (EPC)
One character of the palatine synapomorphous for all the
species of the Curimatidae was previously discussed in detail
by Vari (1983:26). The most common condition of the palatine
in characiform outgroups has a single anterodorsal cartilagi-
nous articular surface contacting the vomerine region of the
neurocranium and the ethmo-palatine cartilage associated with
the upper arm of the premaxilla (the submaxillary cartilage of
Daget, 1964, fig. 23). All species of the Curimatidae differ from
that hypothesized plesiomorphous arrangement in having an
additional, distinct cartilage-capped process on the posterodor-
sal surface of the bone (PDAS, Figure 40A,B). That cartilage
articulates with a corresponding process on the anteroventral
surface of the ventral wing of the lateral ethmoid (Vari, 1983,
fig. 25). Although a second articular surface on the palatine
somewhat comparable to that in the Curimatidae is approxi-
mated to varying degrees in some characiform outgroups, both
in the Neotropical and Ethiopian realms, such a structure is
absent in the proximate outgroups to the Curimatidae, the
families Prochilodontidae, Anostomidae, and Chilodontidae.
The absence of a cartilage capped-articular surface contacting
the lateral ethmoid in those taxa in conjunction with the
morphological differences in the form of similar processes in
those characiform outgroups with such modifications is
congruent with the hypothesis that the posterodorsal articular
process of the palatine characteristic of the Curimatidae
(SYNAPOMORPHY 21) is derived for that family (see Vari,
1983:26, for a more detailed discussion).
Within the Curimatidae, the palatine shows a second set of
modifications that are common to Potamorhina, Curimata,
Psectrogaster, Steindachnerina, Pseudocurimata, Curimatella,
and Cyphocharax. All species of Curimatopsis (Figure 40A)
have a continuous elaborate cartilage that extends along the
anterior face of the palatine (PAL), continues along the
dorsomedial margin of the bone, and expands posterodorsally
into the distinct secondary articular process (PDAS) described
above. The cartilage then continues posteroventrally along the
lateral surface of the palatine arch in the region of contact of
the metapterygoid and ectopterygoid. In Potamorhina, Psec-
trogaster, Curimata, Steindachnerina, Pseudocurimata, Curi-
matella, and Cyphocharax the continuous palatine cartilage
characteristic of Curimatopsis is subdivided into two discrete
portions separated by a well ossified posteromedial region of
the palatine (Figure 40B) . Anteriorly and laterally the anterior
cartilage articulates with the ethmo-palatine cartilage (EPC)
and the vomer respectively. That cartilage is also characterized
by a distinct central depression (Figure 40B) that is not present
in Curimatopsis. The posterodorsal palatine articular cartilage
(PDAS) is a discrete mass enveloped anteriorly, medially, and
to varying degrees laterally by the ossified surface layer of the
palatine. Ventrolaterally the posterolateral articular surface
retains its continuity with the cartilaginous band overlying the
junction between the metapterygoid and the ectopterygoid
44 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
PMX PAL
EPC
PDAS
PAL
B
PMX MET
EPC
FIGURE 40.?Palatine, ethmo-palatine cartilage and proximate portions of
premaxilla and metapterygoid: A, Curimatopsis microlepis, USNM 268867;
and B, Steindachnerina argentea, USNM 285663. (Right side, dorsal view,
anterior to left, dense patterned stippling represents cartilage.)
(Figure 40B) . This subdivision of the palatine cartilage into
two articular surfaces and the associated complex restructuring
of the ossified portions of that element are unique to this
assemblage of genera among examined characiforms and
considered a synapomorphy for Potamorhina, Curimata,
Psectrogaster, Steindachnerina, Pseudocurimata, Curimatella,
and Cyphocharax (SYNAPOMORPHY 62).
The ethmo-palatine cartilage (EPC) is a distinct body
situated between the posterior surface of the upper arm of the
premaxilla and the anterior margin of the palatine. There is
usually only a single cartilage mass lying between these bones,
although that cartilage is apparently subdivided into two units
in some taxa (e.g., the Old World characiform family
Citharinidae, Vari, 1979:294). The ethmo-palatine cartilage is
usually a relatively well developed thick body extending along
most of the anterior margin of the palatine and much of the
rear of the dorsal process of the premaxilla (e.g., Steindachner-
ina argentea, Figure 4 0 B ) to which it is relatively closely joined
by connective tissue bands. Such a condition is found in the
genera Potamorhina, Psectrogaster, Curimata, Steindachner-
ina, Pseudocurimata, Curimatella, and Cyphocharax. In the
species of Curimatopsis, in contrast, the ethmo-palatine
cartilage is reduced to a very small ovoid body along the
posteromedial surface of tip of the premaxilla (Figure 40A).
That condition represents a significant reduction in the relative
size of the cartilage that is considered a synapomorphy for the
members of the genus (SYNAPOMORPHY 49).
Ectopterygoid (ECT)
The ectopterygoid shows a wide range in overall morphol-
ogy within the Curimatidae, although most of the variation
cannot be non-arbitrarily apportioned into discrete characters.
One distinct difference, distinguishing Curimatopsis from the
remainder of the family, involves the form of the anterior
margin of the ectopterygoid. That region of the bone is
unelaborated and transversely rounded in the species of
Curimatopsis but has a distinct flattened vertical flange in
Potamorhina, Curimata, Psectrogaster, Steindachnerina, Pseu-
docurimata, Curimatella, and Cyphocharax. The condition of
the elements in Curimatopsis is similar to that in most
characiforms, and within the context of such an outgroup
analysis would be evaluated as being plesiomorphous. The
Prochilodontidae, the sister group to the Curimatidae, however,
has a highly modified ectopterygoid that is distinctly flattened
transversely. It is thus possible that the flange along the anterior
margin of the bone in non-Curimatopsis curimatids is at least
in part homologous with the anterior portion of the flattened
prochilodontid ectopterygoid. If that is the case, then the
Curimatopsis form of the ossifications would be a derived
secondary reversal to the more generalized condition. Both
hypotheses, that the Curimatopsis form of the ectopterygoid
is derived, or that it is plesiomorphous, would be congruent
with the proposed overall most parsimonious phytogeny. Thus
it is not possible to polarize the character by either outgroup
comparisons or by reference to the overall most parsimonious
hypothesis of intrafamilial relationships.
Mesopterygoid (MES)
The mesopterygoid is typically a relatively simple
ossification situated between the palatine and metapterygoid,
and lying dorsal of the quadrate (Figure 41A). The relative size
of the element varies in different groups of characiforms, with
the condition in the proximate outgroups to the Curimatidae
being a small to moderate sized bone that is unelaborated along
its dorsal margin. The mesopterygoid and metapterygoid which
lie posterior to it along the palatine arch, are suspended from
the neurocranium by a broad thin connective tissue sheet that
attaches along the length of the dorsal margins of those
ossifications.
The described suspensory system joining the mesopterygoid
to the cranium is modified in all curimatids. Members of that
family rather than retaining a broad connective tissue band
between those regions, have instead a discrete ligamentous
band that extends between the dorsal margin of the mesop-
terygoid and the ventral surface of the vomer. Within the
NUMBER 471 45
Curimatidae, this ligamentous band has two different shapes
and forms of attachment to the mesopterygoid. In Curimatop-
sis, the ligamentous band is quite narrow dorsally, but
progressively fans out ventrally to attach over a relatively broad
region of the dorsal margin of the mesopterygoid. In
Potamorhina, Curimata, Psectrogaster, Steindachnerina, Pseu-
docurimata, Curimatella, and Cyphocharax the ligament arises
in a similar fashion from the vomer, but does not fan out
ventrally, being rather more strap-like and of continuous width
from the vomer to the mesopterygoid and attaching onto the
latter element on a distinct dorsal process (AI, Figure 41A,B).
Thickenings within the broad connective tissue sheet
between the palatine arch and neurocranium standard for most
characiforms are found in various taxa, including the genus
Ichthyoelephas in the Prochilodontidae. Those ligamentous
bands, however, are not as discrete as those within the
Curimatidae, nor do they have the same areas of attachment.
Thus those structures are considered non-homologous with the
discrete ligament between the vomer and mesopterygoid
characteristic of the Curimatidae. Given the broad area of
attachment of the connective tissue sheet typically joining the
palatine arch and the neurocranium in characiforms, these
reductions in the area of attachment of the connective tissues
extending between the vomer and mesopterygoid in curimatids
are treated as derived.
In the Curimatidae, such reductions in the area of attachment
are considered apomorphous at two levels of universality. The
possession of a fan-shaped or strap-like thickened connective
tissue band arising from a restricted area on the vomer and
having either a broad or narrow attachment on the mesop-
terygoid is hypothesized a synapomorphy for all members of
the Curimatidae (SYNAPOMORPHY 34). The presence of a
strap-like ligament of constant width with a discrete further
restricted area of attachment to the mesopterygoid, in
comparison with the more extensive area of attachment in
Curimatopsis and characiform outgroups, is, in turn, hypothe-
sized to be a shared derived character for Potamorhina,
Curimata, Psectrogaster, Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax (SYNAPOMORPHY 63).
The mesopterygoid in characiforms most commonly lies
dorsal of the junction of the posterodorsal portion of the
quadrate and the anteroventral region of the metapterygoid,
with those elements meeting along a synchondral joint below
the mesopterygoid. A majority of characiforms have a
moderately to well developed aperture, the metapterygoid-
quadrate fenestra (MQF, Figure 41A) , delimited anteriorly and
ventrally by the quadrate (QU), dorsally and posteriorly by the
metapterygoid (MET) and sometimes posteroventrally by the
symplectic (SYM) (e.g., Brycon meeki, Weitzman, 1964, fig.
10). Such an association of bones and cartilages is considered
plesiomorphous in light of comparisons to both proximate
outgroups to the Curimatidae and diverse other characiforms.
Within the Curimatidae that pattern of bones also occurs in the
species of Curimatopsis and Potamorhina (Figure 41A). The
PAL
MES
MET
MQF
SYM
QU PSMQF
ASMQF
FIGURE 41.?Mesopterygoid: A, Potamorhina laticeps, USNM 121325; and
B, Psectrogaster ciliata, USNM 26990. (Right side, medial view, anterior to
left, proximate portions of neighboring bones outlined and lightly stippled.)
species of Psectrogaster, Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax have the mesopterygoid
significantly modified. In those genera, the mesopterygoid no
longer terminates ventrally distinctly short of the dorsal margin
of the metapterygoid-quadrate fenestra. Rather the mesop-
terygoid is further expanded ventrally over the medial surface
of the junction of the metapterygoid and quadrate and continues
partially or completely across the anterior portion of the
metapterygoid-quadrate fenestra (Figure 41B). This extension
of the mesopyterygoid results in the subdivision of that fenestra
to varying degrees into anterior (ASMQF) and posterior
(PSMQF) sections. Such an elaboration of the mesopterygoid,
and the assocated subdivision of the metapterygoid-quadrate
fenestra has not been encountered in any examined characiform
outgroups. That ventral expansion of the mesopterygoid is
consequently considered a synapomorphy for the clade
consisting of Psectrogaster, Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax (SYNAPOMORPHY 85).
Metapterygoid (MET)
Van (1983:28) noted that members of the Curimatidae have
an apomorphous elaboration of the metapterygoid consisting
of a ridge extending along the medial surface of the bone for
at least part of its length. That condition, hypothesized to be a
synapomorphy for the members of the Curimatidae (SYNAPO-
MORPHY 23), is further developed in various sublineages of the
46 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
family. The ridge (METR) is only moderately developed in
Curimatopsis (Figure 42A) and Potamorhina (Figure 42B), but
is much more pronounced in Curimata (Figure 42D),
Psectrogaster (Figure 42c), Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax in which it extends across the
entire length of the bone. This more extensive development
of the metapterygoid ridge is hypothesized to be a more derived
condition in light of the absence of any form of the ridge in
examined outgroups and relative to the less developed ridge
in Curimatopsis and Potamorhina. The possession of such a
process is thus considered a synapomorphy for Curimata,
Psectrogaster, Steindachnerina, Pseudocurimata, Curimatella,
and Cyphocharax (SYNAPOMORPHY 77).
The species of Curimata have the ridge on the medial surface
of the metapterygoid considerably thickened anteriorly (Figure
42D) . This expanded region of the bone is continuous anteriorly
with the comparably expanded region on the posteromedial
portion of the mesopterygoid that serves as the area of
attachment for the ligament extending between that element
and the vomer. This expansion of the anterior portion of the
metapterygoid, unique to the species of Curimata, is hypothe-
sized synapomorphous for the genus (SYNAPOMORPHY 83).
As noted in the discussion of the mesopterygoid in the
previous section, the metapterygoid and quadrate jointly
delimit a usually relatively large, rotund, or horizontally oblong
aperture in the middle portion of the suspensorium. That
opening, the metapterygoid-quadrate fenestra, is widely
distributed within the Characiformes, being found in the
Neotropical Characidae (Brycon, Weitzman, 1962:46, fig. 10;
Acestrorhynchus Eigenmann and Kennedy, Roberts, 1969, fig.
32), Ctenoluciidae {Ctenolucius Gill, Roberts, 1969, fig. 28),
Erythrinidae (Hoplias, Roberts, 1969, fig. 29), Hemiodontidae
(Hemiodus, Roberts, 1974, fig. 7), Parodontidae, Prochilodon-
tidae (Ichthyoelephas, Vari, 1983, fig. 28), and Anostomidae
(Winterbottom, 1980, fig. 35), and in the Old World families
Characidae (Alestes, Hydrocynus [Brewster, 1986, fig. 8], and
Phenacogrammus), Hepsetidae (Hepsetus Swainson), Cithar-
inidae (Citharinus, Daget, 1962a, fig. 9; Citharidium, Daget,
1962b, fig. 13), and most members of the Distichodontidae
(Xenocharax, Daget, 1960, fig. 10). Although this fenestra is
subdivided by a posteroventral process of the mesopterygoid
in most genera of the Curimatidae (see previous section), the
overall form of opening characteristic of most characiforms is
found in all curimatids other than the species of Curimatopsis.
In that genus the fenestra is nearly eradicated by the ventral
expansion of the metapterygoid which extends nearly to the
MET METR
MES
MET
MQF
SYM
SYM
FIGURE 42.?Metapterygoid: A, Curimatopsis microlepis, USNM 268867; B, Potamorhina allamazonica, USNM
257367; c, Pseclrogasler ciliata, USNM 269990; and D) Curimata cyprinoides, USNM 267964. (Right side,
medial view, anterior to left, proximate portions of neighboring bones outlined and lightly stippled.)
NUMBER 471 47
dorsal margin of the posterior portion of the quadrate (Figure
42A). This reduction in the extent of the aperture differs from
the diminished opening in some members of the Old World
characiform family Distichodontidae in the mode in which it
is achieved. In the distichodontid Neolebias spilotaenia
Boulenger the reduced fenestra is a consequence of the
expanded symplectic rather than an enlarged metapterygoid
(Vari, 1979:293). Nannocharax Giinther and Hemigrammo-
charax Pellegrin of the same family (see Daget, 1961, fig. 10),
along with members of the Neotropical family Lebiasinidae
(see Weitzman, 1964, fig. 7) have the quadrate approximating
the metatperygoid as a consequence of the reduced vertical
extent of the entire suspensorium rather than due to an
expansion of individual elements. The conditions in those
outgroup families thus all differ from that found in Curimatop-
sis.
A near or total elimination of the fenestra by an expanded
metapterygoid is found, however, in all members of the
Chilodontidae. The closest relatives to the Chilodontidae, the
family Anostomidae is, in contrast, characterized by the
presence of a fenestra (see Winterbottom, 1980, fig. 35).
Similarly the Prochilodontidae, the sister group to the
Curimatidae, has a complete metapterygoid-quadrate fenestra
typical of characiform outgroups (e.g., Ichthyoelephas, Vari,
1983, fig. 28) as do all curimatids other than Curimatopsis. It
is thus most parsimonious to assume that the conditions in the
Chilodontidae and Curimatopsis are homoplasious. The
Curimatopsis condition of a reduction of the metapterygoid-
quadrate fenestra as a consequence of the expansion of the
metapterygoid is thus considered a synapomorphy for the
members of the genus (SYNAPOMORPHY 39).
SUPRANEURALS ( S N j , ^
A series of midsagittal ossifications are typically found
dorsal of the vertebral column in the region between the rear
of the neurocranium and the anteriormost proximal radial
pterygiophore of the dorsal fin. These bones, the supraneurals,
along with the proximal radial pterygiophores of the dorsal fin
interdigitate basally between the distal portions of the neural
spines of the proximate vertebrae. In the following discussion
and in Figure 43 neural spines are numbered sequentially from
the spine of the first free vertebrae posterior of the Weberian
complex.
The typical pattern for these elements in the Curimatidae is
that shown in Figure 43A. TWO supraneurals are located
anterior of the first neural spines (the two anterior supraneurals
are sometimes fused), with the third to fifth supraneurals and
the second to fourth neural spines interdigitating sequentially.
The basal portion of the first proximal radial pterygiophore of
the dorsal fin interdigitates between the distal portions of the
fourth and fifth neural spines. This pattern of supraneurals,
neural spines, and the first proximal pterygiophore is
generalized for the vast majority of curimatid species with three
exceptions. Two nominal species in Curimata (yittata and
murieli) have the number of supraneurals reduced to four. This
reduction is apparently a consequence of the loss of the
posterior supraneural, and as a consequence the most posterior
element of that series in these species is located between the
second and third neural spines. Associated with that reduction
in the number of supraneurals is the shift anteriorly of the
region of interdigitation of the basal portion of the first
proximal pterygiophore of the dorsal fin to between the third
and fourth neural spines, rather than between the fifth and sixth
spines.
The species of Pseudocurimata have the opposite modifica-
tion of the pattern of association of the supraneurals and the
anterior proximal radial pterygiophore with the anterior neural
spines. In that genus the anterior proximal pterygiophore of the
dorsal fin inserts between the distal portions of the fifth and
sixth, or sixth and seventh neural spines rather than between
the fourth and fifth neural spines as in the vast majority of
curimatids. Associated with that more posterior insertion of the
fin is the absence of a supraneural or proximal pterygiophore
either between the fourth and fifth, or fifth and sixth neural
spines (Figure 43B,C).
Outgroup information on the polarity of the different
FIGURE 43.?Supraneurals, neural spines of proximate vertebrae, and first
proximal radial pterygiophore of dorsal fin: A, Curimata cyprinoides, USNM
225619; B, Psuedocurimata lineopunctata, MCZ 54029; and C, Pseudocuri-
mata peruana, USNM 285667. (Left side, lateral view, anterior to left.)
48 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
patterns of interdigitation is equivocal. Various species in the
Prochilodontidae have the first proximal pterygiophore of the
dorsal fin inserting either between the fourth and fifth or fifth
and sixth neural spines, and demonstrate overall patterns of
supraneurals, neural spines, and proximal pterygiophores
comparable to those shown in Figure 43A,B. Members of the
Anostomidae also show some variation in the interdigitation
pattern of these elements, whereas the species of the
Chilodontidae have the first proximal pterygiophore inserting
between the fourth and fifth neural spines, with the overall
pattern of that element, the neural spines, and supraneurals
similar to the pattern of Figure 43A. Given the distribution of
the various states within the Curimatidae and in its sister groups
of varying degrees of proximity, it is most parsimonious to
hypothesize that the position of the first proximal pterygio-
phore of the dorsal fin posterior of the fifth or sixth neural
spines in Pseudocurimata is derived (SYNAPOMORPHY 102).
Such a posterior portion of the dorsal fin and the resultant
gap in the sequence of interdigitations of elements between the
neural spines immediately anterior to the first proximal
pterygiophore of the dorsal fin that is general to all
Pseudocurimata species also occurs in two other nominal
species within the Curimatidae. Curimata ocellata and C.
semitaeniata, elongate fusiform species of the Amazon basin
have a pattern of supraneurals and the anterior proximal
pterygiophore of the dorsal fin comparable to that of some
Pseudocurimata species (e.g., Pseudocurimata lineopunctata,
Figure 43B) . The common occurrence of this derived pattern
in Curimata ocellata and C. semitaeniata on the one hand, and
Pseudocurimata species on the other is considered homopla-
sious within the context of the overall most parsimonious
hypothesis of relationships.
CAUDAL FIN COMPLEX
The hypural complex and the rays of the caudal fin
demonstrate a number of apomorphous modifications in
Curimatopsis which were previously discussed and illustrated
by Van (l982a:5-9). Four of those characters involve sexually
dimorphic features in which the derived condition is found
only in the males. These apomorphies are the expansion of the
posterior neural and haemal spines on the posterior two free
centra into broad midsaggital plates (see Vari, 1982a, figs. 2,
4; SYNAPOMORPHY 35), the possession of basal spurs on the
middle caudal-fin rays (see Vari, 1982a, fig. 5; SYNAPOMOR-
PHY 36), the vertical expansion of the penultimate ray of the
lower lobe of the caudal fin (see Vari, 1982a, fig. 5;
SYNAPOMORPHY 37), and the close association of the
ventralmost ray of the upper lobe of the caudal fin with the
dorsalmost ray of the ventral lobe of that fin rather than the
retention of the gap between those elements typical for
characiforms (SYNAPOMORPHY 38). Further data on those
characters and the conditions in those systems in outgroup taxa
are found in Vari (1982a:5-9). Three other derived features
of the caudal skeleton in Curimatopsis, the pronounced
expansion of the penultimate principal caudal-fin ray of males
(Curimatopsis SYNAPOMORPHY C), the possession of a second
hypural free from PUj +Ul but fused to the first hypural
(Curimatopsis SYNAPOMORPHY D), and the contact in males
of the third hypural with the fused first and second hypurals
and the associated development posteriorly of interdigitations
between those elements (Curimatopsis SYNAPOMORPHY E),
were advanced as synapomorphies of the clade consisting of
C. crypticus and C. evelynae. Curimatopsis myersi described
subsequendy (Vari, 1982b) has been found to also share these
apomorphous features, and the more inclusive outgroup
examinations associated with this study has confirmed the
derived nature of the characters.
An additional synapomorphy in the caudal skeleton pertinent
to hypotheses of subgeneric relationships in Curimatopsis has
also been uncovered in this study. In males of C. evelynae, C.
crypticus, and C. myersi, the posterior two free preural centra
and the fused Pl^ + Uj have distinct longitudinal dorsoven-
trally flattened lateral flanges arising from their midlateral
surface. The processes evidently extend into the connective
tissue layer between the epaxial and hypaxial musculature, and
serve an an area of attachment for various ligamentous bands.
Such processes on those posterior vertebrae are unknown
elsewhere in the Curimatidae, nor have they been encountered
in the examined outgroups. The lateral flanges on these
posterior vertebral elements are thus hypothesized to be
synapomorphies for Curimatopsis evelynae, C. crypticus, and
C. meyersi (Curimatopsis SYNAPOMORPHY G).
A reduction to one set of uroneurals from the two sets typical
for characiforms was hypothesized to be a synapomorphy for
the species of Curimatopsis (Vari, 1982a:4,5; SYNAPOMORPHY
40), a polarity congruent with the findings of this project.
Further analysis has shown that a greatly reduced or absent set
of second uroneurals is also characteristic of all species of
Pseudocurimata. Within an examined series of six cleared and
stained individuals of P. boulengeri, three were found to lack
any indication of a second set of uroneurals and in the other
three specimens there were present only very small bones on
each side of the hypural complex in the region typically
occupied by the second uroneurals. No indication was found
of the second set of uroneurals in the examined cleared and
stained specimens of P. peruanus or P. troscheli, and only a
greatly reduced set of second uroneurals was found in the
examined cleared individuals of P. lineopunctatus and P.
patiae. This pronounced reduction or complete loss of the
second set of uroneurals is hypothesized synapomorphous for
the members of Pseudocurimata (SYNAPOMORPHY 103). That
character is considered to be homoplasiously present in
Curimatopsis within the context of the overall most parsimoni-
ous hypothesis of relationships within the Curimatidae.
NUMBER 471 49
LATEROSENSORY SYSTEM OF BODY
The vast majority of characiforms have a completely
developed laterosensory canal system on the body with all
lateral line scales having minimally a pore opening to the
surface, and with some species and groups having more
elaborate canal systems within the scales. A completely pored
lateral line occurs in all members of the proximate outgroups
to the Curimatidae, the families Prochilodontidae, Anostomi-
dae, and Chilodontidae. Within the Curimatidae a completely
pored laterosensory system on the body characterizes all
members of the Curimatidae with the exception of the species
of Curimatopsis and a subset of Cyphocharax species
(saladensis, vanderi, punctata). Curimatopsis maulatus, de-
scribed by Ahl (1934) as having an incompletely pored
laterosensory canal system, is actually a juvenile of a
Cyphocharax species in which that system is fully developed
in the adults (pers. observ.). Although the reduction in this
poring is hypothesized as derived given the fully developed
laterosensory canal systems of the outgroups, the distribution
of that character is not congruent with the final phylogeny of
this study (Figure 44). The character is thus considered
homoplasious at the generic level, with the reduced develop-
ment of poring of the lateral line hypothesized synapomor-
phous for Curimatopsis species on the one hand (SYNAPOMOR-
PHY 50) and the cited Cyphocharax species on the other.
A possible correlation exists between the reduced degree of
poring of the laterosensory canal system of the body and the
relatively small body size of many of the species with that
character. Although paedomorphic features including a reduc-
tion in the degree of the development of the laterosensory canal
system are often associated with relative miniaturization
(Weitzman and Vari, 1988), the concept of such a direct
correlation would appear at first to be qualified when applied
to the Curimatidae since Curimatopsis microlepis, a species
which achieves at least 90 mm SL, has an incompletely pored
lateral-line. The concept of a miniature and the correlated
suggestion of paedomorphosis is, nonetheless, applicable even
to C. microlepis given the much larger body sizes of all
members of the Prochilodontidae, the sister group to the
Curimatidae, and proximate sister groups to Curimatopsis
within the Curimatidae (see also discussion under "Convergent
Characters").
SQUAMATION
The species of the Curimatidae are characterized by
complete squamation of the body, but without basal sheaths
of scales on the dorsal or anal fin such as occur on the anal fin
in the characid Markiana Eigenmann. Similarly the squamation
on the caudal fin terminates only slighty posterior of the base
of the caudal fin rays in the vast majority of the members of
the Curimatidae. The one exception involves the species of
Curimatella. Although juveniles of the genus have only a
limited degree of squamation at the base of the caudal fin
comparable to the situation in curimatid outgroups, there is an
ontogenetic increase in the extent of squamation, with adults
having scales over much of the proximate two-thirds of both
lobes of the caudal fin.
The absence of such an extensive patch of scales on the
caudal fin lobes, the condition in the remainder of the
Curimatidae, is also characteristic of the Prochilodontidae and
Chilodontidae, along with all genera of the Anostomidae other
than for the few species of the genus Leporellus Liitken. Within
the context of the overall phylogenetic hypothesis for the clade
consisting of the Curimatidae, Prochilodontidae, Anostomidae,
and Chilodontidae (Vari, 1983), the presence of caudal fin
squamation in Curimatella is most simply considered synapo-
morphous for the members of the genus (SYNAPOMORPHY
104), albeit independently derived within the Anostomidae.
The relative size of body scales within the family
Curimatidae as reflected in the number of scales along the
lateral line from the supracleithrum to the base of the caudal
fin demonstrates a nearly four fold range. That variation forms,
however, a near continuum that makes it impossible to
non-arbitrarily utilize much of that data for phylogenetic
studies. Two exceptions involve the very high lateral-line
counts for the species oiPotamorhina, along with Cyphocharax
abramoides. Vari (1984a:7) noted that the species of Pota-
morhina have 85-110 scales in that series, overlapping within
the Curimatidae only with Cyphocharax abramoides (77-95
lateral line scales). The remainder of the family have 26 to 76
pored scales in a longitudinal series, and the Prochilodontidae,
the sister group to the Curimatidae, typically has less that 80
lateral-line scales (R.M.C. Castro, pers. comm.). Thus the
condition in Potamorhina is judged to be synapomorphous for
the species of the genus (SYNAPOMORPHY 69) (see Vari,
1984a:7 for more details). The increased, although not as
numerous, lateral line scales of Cyphocharax abramoides are
considered to represent a homoplasious increase in that series.
Synapomorphy List and Phylogenetic Reconstruction
The previous section of the discussion details the numerous
derived modifications of and within the Curimatidae over a
variety of body systems that were discovered in the course of
this and previous studies (Vari, 1982a, 1983, 1984a). Vari
(1983:48) hypothesized that the Curimatidae and Prochilodon-
tidae were sister groups, and that the Curimatidae (not
including Anodus Spix and Eigenmannina Fowler) was a
monophyletic assemblage. The anatomical studies associated
with this project have uncovered additional information
congruent with both of those hypotheses. The data from the
examined body systems is also both relevant to a hypothesis
of generic level relationships within that family, and to the
evaluation of previously proposed genera.
50 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
In the following, shared derived characters congruent with
a hypothesis of monophyly of the group consisting of the
Prochilodontidae and Curimatidae will be discussed first. The
majority of the characters supporting the hypothesis of the
monophyly of that bifamilial lineage were analyzed and listed
by Van (1983) and all are noted again below. Some of these
were not redescribed in detail in the preceeding sections of this
paper, and the reader is referred to Van (1983) for detailed
information on those characters. Additional synapomorphies
for the Curimatidae and Prochilodontidae were discussed in
preceeding character descriptions, and are also listed below.
The enumeration of the synapomorphies for the clade formed
by those two families is followed by a listing and discussion
of the less universal apomorphous states that characterize the
Curimatidae, and the clades, named and unnamed, within that
family.
A variety of alternative schemes have been proposed to
convey phylogenetic information within a classification. A
number of these systems involve the proposal of a name for
each clade defined by shared derived characters, a method that
would necessitate a number of taxa between the generic and
familial levels in the Curimatidae. Formal nomenclatural
recognition is given in this study only to the family and the
major terminal sublineages recognized as genera. This
simplified nomenclatural system is used since the recognition
of taxa between the generic and familial levels for those
subfamilial clades consisting of two or more genera would
necessitate the proposal and utilization of at least four
suprageneric taxa, a number that would increase if the terminal
polychotomy were to be resolved by future studies. The
resultant proliferation of suprageneric taxa within the family
would do little to clarify the following discussion, or indeed
might obfuscate the presentation. Although the proposed
scheme of relationships cannot as a consequence be directly
CURIMATIDAE
-1-15
FIGURE 44.?Cladogram of the most parsimonious hypothesis of phylogenetic relationships between the families
Prochilodontidae and Curimatidae and within the Curimatidae (see Van, 1983, for a discussion of
synapomorphies for the Prochilodontidae). Numbered synapomorphies correspond to text.
NUMBER 471 51
retrieved from the classification, that hypothesis is discussed
in detail in the following section and is presented visually in
Figure 44.
Characters pertinent to phylogenetic reconstructions at the
subgeneric levels are not typically be discussed unless they
represent homoplasies found at both the generic and/or
suprageneric levels on the one hand, and at subgeneric levels
on the other. Exceptions are made for those unique synapomor-
phies that are pertinent to the hypotheses of species level
relationships advanced by Vari for Curimatopsis (1982a) and
Potamorhina (1984a), and which are in some instances
reiterated or even expanded in this study.
Subsequent to the reconstruction of the most parsimonious
hypothesis of phylogenetic relationships among the genera
within the Curimatidae, I discuss the distribution of homopla-
sious characters, those derived attributes with an incongruent
distribution on the cladogram (Figure 44). The discussion deals
with the two types of homoplasies indicated by the final
phylogeny. The first class involves derived characters that are
common to two subunits of the family which are not each
other's closest relatives. The second group consists of evidently
identical derived characters that occur both within a subunit
of the Curimatidae and in some characiform outgroup. The
discussion of the homoplasies in conjunction with the
phylogenetic reconstruction provides the basis for the evalu-
ation of the efficacy of previous classificatory schemes as
indicators of the phylogenetic history of the components of the
family, that is its suprageneric groupings and genera as
recognized in this study. Convergent characters are also
evaluated in terms of whether they represent innovative or
reductive attributes, and relative to the degree of congruence
in those attributes within subunits of the Curimatidae.
The most parsimonious hypothesis incorporating the previ-
ously described synapomorphies, including homoplasies, is
presented in Figure 44. This hypothesis was first derived by
manual construction. Its parsimony was confirmed by use of
David L. Swofford's (1985) numeric computer algorithm
PAUP (Phylogenetic Analysis Using Parsimony), version 2.4.
Synapomorphies unique to lineages recognized as genera in
this study were not included in the analysis except in the case
of genera with components that shared derived features with
another generic or suprageneric clade within the Curimatidae.
The data on 49 discretely variable polarized features with
two to four character states was analyzed using the branch-
andbound option of PAUP, guaranteed to find the most
parsimonious trees. The analysis resulted in fifteen equally
parsimonious trees with consistency indices of 0.825. These
trees had identical topologies other than for the sequence of
the genera Curimatella, Cyphocharax, Steindachnerina, and
Pseudocurimata. Since those taxa form a terminal multicho-
tomy, these fifteen trees are identical in terms of the
phylogenetic hypothesis that they represent, and equivalent to
the scheme of relationships in Figure 44. Characters are
numbered sequentially from earlier to later dichotomies,
simplifying the vizualization of character distribution, and
allowing cross-referencinge of character descriptions and text
discussions. The numbering of characters in the following text
and in "Character Description and Analysis" corresponds to
the numbered synapomorphies of Figure 44.
Characters pertinent to the species level phylogenetic
reconstruction in Curimatopsis are sequentially identified
alphabetically within the discussions of that genus and in
Figure 45, but are not included in Figure 44. These form a
separate character set.
MONOPHYLY OF THE CURIMATIDAE
AND P R O C H I L O D O N T I D A E CLADE
Vari (1983:47) advanced eleven synapomorphies supporting
the hypothesis that the Prochilodontidae is the sister group to
the Curimatidae. These are listed below. Numbers 3,9,10, and
11 were discussed in detail by Vari (1983), to which the reader
is referred for additional information. For the remaining
synapomorphies, additional description and outgroup compari-
son may be found herein under "Character Description and
Analysis."
1. Reorientation of the dorsal process of the fourth epibran-
chial anteriorly with its resultant extension over the dorsal
surface of the fourth infrapharyngobranchial (Figures
5-13).
2. Anterodorsal expansion of the cartilaginous fifth epibran-
chial, its attachment to the posterodorsal margin of the
fourth epibranchial, and the resultant encirclement by
those elements of the fifth efferent branchial artery
(Figures 5,6, 8-11).
3. Large sac-like muscular epibranchial organ that extends
dorsal to the medial elements of the dorsal portions of the
gill arches (Vari, 1983:21-24).
4. Conversion of the plesiomorphously flat fourth upper
pharyngeal tooth plate into a curved ossification that is
wrapped around the ventral, lateral, and medial surfaces
of the fourth infrapharyngobranchial (Figures 5, 6, 8-16).
5. Reduction or loss of dentition on the ventral surface of the
fifth upper pharygneal tooth plate (Figures 5, 7-11).
6. Reduction or loss of an ossified first basibranchial.
7. Absence of dentition on the dorsal surface of the fifth
ceratobranchial (Figures 25, 27).
8. Anteromedially directed process on the ventral surface of
the fourth ceratobranchial.
9. Distinct posteroventrally aligned flange on the lateral
surface of the opercle or a further derived condition of
that process (Vari, 1983:29-30).
10. Increase in the depth and width of the fossa for the
scaphium resulting in the interconnections of the fossa
with the lateral occipital foramen laterally, and the interior
of the cranium anteriorly (Vari, 1983:38-41, fig. 34).
11. Posterior development of the lateral margin of the
52 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
exoccipital lateral to the foramen magnum, thereby
forming a common aperture for the foramen magnum,
cavum sinus imparis, and paired fossae for the scaphium.
The exoccipital consequently forms a cover laterally for
the the anterolateral surface of the scaphium (Vari,
1983:38^1, fig. 33).
The following additional derived characters shared by and
uniting the Curimatidae and Prochilodontidae were discovered
during this study. Additional description can be found under
relevant sections of "Character Description and Analysis."
12. Constriction of the complex formed by the fourth
infrapharyngobranchial and fourth upper pharyngeal tooth
plate, resulting in a convex lateral margin to those-
conjoined elements (Figures 15, 16).
13. Expansion and restructuring of the uncinate process of the
third epibranchial into an anterior process overlying the
line of articulation between the third infrapharyngobranchial
and third epibranchial (Figure 19).
14. Subdivision of the anterior articular surface of the third
epibranchial, with the expansion and reorientation of the
medial portion of the articulation into a medially directed
ossified process extending along a matching groove on the
the anterodorsal portion of the fourth infrapharyngobran-
chial (Figures 5, 6, 8-11,19).
15. Expansion of the posterolateral portion of the third
infrapharyngobranchial into a distinct dorsal flange
(Figures 5, 6, 8-11).
As noted by Vari (1983:54-55) the phylogenetic hypothesis
of a sister group relationship between the Curimatidae and
Prochilodontidae agrees, partly or totally, with the classifica-
tions proposed by Giinther (1864), Boulenger (1904), Regan
(1911), Eigenmann (1917), Gregory and Conrad (1938), and
Gery (1977b). All those authors included the majority of the
taxa that are today considered to form the Curimatidae and
Prochilodontidae in more encompassing groupings, along with
taxa with which they are not closely related (Vari, 1983:54-55
for a more indepth discussion). Roberts (1973), in a significant
shift from those earlier classifications, proposed that the
relationships of the Prochilodontidae lies with the Anostomidae
or the lineage formed by the Anostomidae and Chilodontidae
rather than with the Curimatidae. Synapomorphies 12 to 15
discovered during this study, provide additional support for the
hypothesis of the sister group relationship between the
Curimatidae and Prochilodontidae. Those synapomorphies
along with the shared derived characters proposed previously
by Vari (1983) (SYNAPOMORPHIES 1-11) are incongruent with
Roberts' (1973) hypothesis of a closer phylogenetic association
of the Prochilodontidae with the Anostomidae, or with the
group formed by the Anostomidae and Childontidae rather
than with the Curimatidae.
MONOPHYLY AND INTRAFAMILIAL RELATIONSHIPS
OF THE FAMILY CURIMATIDAE
In a previous publication Vari (1983:48) noted that the
Curimatidae was the least derived family among the taxa under
discussion (Curimatidae, Prochilodontidae, Anostomidae, and
Chilodontidae) in terms of number of synapomorphies (ten)
found. The ten synapomorphies found in that study (Vari,
1983) for the Curimatidae are listed below. Of these,
synapomorphies 17, 22, 24, and 25 were discussed in detail
by Vari (1983) and are not further detailed within this paper.
Additional discussion of the remaining synapomorphies may
be found under "Character Description and Analysis."
16. Absence of dentition on the dentary and premaxilla.
17. Possession of a third posttemporal fossa bordered solely
by the epioccipital (Vari, 1983:37-38).
18. Expansion of the anteroventral portion of the third
hypobranchial into a vertical flange with anterior and
posterior processes (Figures 23-24).
19. Absence of an ossified first basibranchial (Figure 25).
20. Posteroventral expansion of fifth upper pharyngeal tooth
plate into curved, convoluted process (Figures 5, 7, 11).
21. Secondary posterodorsal, cartilage capped process on the
palatine which contacts a comparable process on the
ventral wing of the lateral ethmoid (Figure 40).
22. Cartilage-capped articular process along the ventral edge
of the ventral wing of the lateral ethmoid that articulates
with the palatine (Vari, 1983:26-27, fig. 25).
23. Horizontal shelf on the medial surface of the metapterygoid.
24. Large cartilaginous ethmoid block and widened mesethmoid
(Vari, 1983, figs. 36, 37).
25. Enlarged lagenar capsule (Vari, 1983:41).
Nine additional synapomorphies for the Curimatidae were
discovered during this study (see also descriptions under
"Character Description and Analysis"). They are as follows:
26. Ventral ridge extending from posterior of anteroventral
articular surface of fourth epibranchial along ventral
surface of the main body of that element (Figure 5).
27. Anterior reorientation of the primary axis of the dorsal
process of the fourth epibranchial beyond the condition
in the Prochilodontidae (Figures 5, 7-11).
28. Longitudinal ridges on the dorsal surface of the fourth
ceratobranchial (Figure 26).
29. Dorsal convexity of the fifth ceratobranchial.
30. Transverse expansion of the fifth ceratobranchial both
medial and lateral of the line between the anterior and
posterolateral cartilage bodies (Figure 27B,C).
31. Presence of a well developed cord-like ligament running
from the ventral hypohyal between the medial branchios-
tegal rays and extending along a groove on the midventral
surface of the urohyal (Figure 32).
32. Extension of the urohyal posteriorly beyond the point
NUMBER 471 53
where the medial branchiostegal rays diverge laterally
(secondarily shortened in Curimata and the clade consist-
ing of Steindachnerina, Pseudocurimata, Curimatella, and
Cyphocharax (see discussion under "Hyoid Arch")).
33. Anterolateral expansion of the portions of the first to third
branchiostegal rays proximate to the hyoid arch, and the
resulting overlap of the articulation of the second
branchiostegal with the anterior ceratohyal by the anterior
portion of the third branchiostegal.
34. Possession of a discrete ligamentous band, with a restricted
area of attachment on the vomer, joining the neurocranium
and the mesopterygoid.
As noted in the discussion of the "Opercular Aparatus," the
form of the opercle and subopercle may represent additional
synapomorphies for the family, but were not included here due
to problems with non-arbitrary quantification .
The genera Anodus and Eigenmannina have traditionally
been placed in the Curimatidae (e.g., Fernandez-Yepez, 1948)
or associated in an unspecified manner with that family (e.g.,
Eigenmann, 1917) as a consequence of their common lack of
jaw dentition. This character, unique to Anodus, Eigenman-
nina, and adults of the Curimatidae among characiforms, both
New and Old World, was presumably the basis for the original
description of three species of curimatids in Anodus {Anodus
latior Spix = Potamorhina latior, Anodus ciliatus Miiller and
Troschel = Psectrogaster ciliata, and Anodus troschelii
Giinther = Pseudocurimata troschelii)- Roberts (1974), in a
significant shift, proposed that Anodus and Eigenmannina
(considered congeneric by Roberts, 1974) were members of
the Hemiodontidae, and not closely related to to Curimatidae.
Subsequently, Gery (1977b:239) commented upon the overall
similarity of hemiodontids, Anodus, and Eigenmannina, and
noted Roberts' (1974) publication in a footnote, but did not
otherwise comment on the relationships of the different taxa.
More recently, G6ry (1984:357) has retained the "Anodins"
(Anodus and Eigenmannina) in a broadly defined Curimatidae
that does not form a monophyletic assemblage, and which he
delimits from his "Hemiodins" (= Hemiodontidae). Although
G6ry (1984) noted that 'The Anodins [Anodus and Eigenman-
nina] bear many resemblances to the Hemiodins," he neither
provided any reason for his rejection of Roberts' (1974)
hypothesis, nor did he elaborate on his continued alignment
of Anodus and Eigenmannina with the Curimatidae.
The attributes listed by Roberts (1974:429) as justification
for the inclusion of Anodus and Eigenmannina in the
Hemiodontidae were an amalgam of three classes of characters.
A number are synapomorphies for the Curimatidae (sensu
stricto) (e.g., the cartilaginous articular surfaces between the
palatine and the lateral ethmoid) that do define that family and
serve to distinguish it from Anodus and Eigenmannina, but
do not provide any insight into the phylogenetic position of
those two genera to either the Curimatidae (sensu stricto) or
the Hemiodontidae. The second group of features are shared
primitive characters common to the Hemiodontidae, Anodus,
and Eigenmannina (e.g., moderate coiling of the intestine), and
which also occur in other groups of characiforms but are found
in a derived condition in the Curimatidae. Such plesiomorphies
are not useful in determining whether these three taxa form a
monophyletic unit, or their possible phyletic associations with
the Curimatidae. The third class of characters are shared
derived modifications common to the Hemiodontidae, Anodus,
and Eigenmannina, but which are not typical of the
Curimatidae. Characters of that class cited by Roberts (1974)
include the increased number of vertebrae, and the fusion of
the first and second hypurals into a single ossification
autogenous from the remainder of the hypural fan. Such
characters are appropriately used as indicators of relationship
and support Roberts' (1974) hypothesis that Anodus and
Eigenmannina are aligned with the Hemiodontidae.
A more indepth evaluation of the placement of Anodus and
Eigenmannina within the Hemiodontidae and a more rigorous
definition of that family based on shared derived characters are
both outside the scope of this study. The results of this study
are, however, congruent with the hypothesis that Anodus and
Eigenmannina are not closely related to the Curimatidae. Those
genera share the synapomorphies noted above indicating that
they are aligned phyletically with the Hemiodontidae. Further-
more those two genera lack the vast majority of the numerous
hypothesized synapomorphies for the clade consisting of the
Prochilodontidae and Curimatidae (SYNAPOMORPHIES 1-15),
and the shared derived characters that delimit the Curimatidae
(SYNAPOMORPHIES 16-34). The continued retention of Anodus
and Eigenmannina in the Curimatidae would consequently
require a hypothesis of secondary loss of nearly all of the
derived characters in those two genera. That factor in
combination with some of the characters cited by Roberts
(1974) (see paragraph above) refute the implicit hypothesis of
a close phyletic relationship of Anodus and Eigenmannina with
the Curimatidae as advanced by G6ry (1984).
The scheme of hypothesized phylogenetic relationships
within the Curimatidae (Figure 44) has a basal pinnate form.
The first dichotomy is between the clade consisting of the five
species of Curimatopsis and the lineage formed by Pota-
morhina, Curimata, Psectrogaster, Steindachnerina, Pseu-
docurimata, Curimatella, and Cyphocharax.
Curimatopsis
Van (1982a), in a revisionary study of Curimatopsis,
advanced eight synapomorphies that served to delimit the four
then-known species in the genus. These shared derived
characters are also common to an additional species, C. myersi,
that was described subsequently (Vari, 1982b). The five species
of Curimatopsis occur in the Orinoco, Amazonas, and
Paraguay rivers, and in some rivers of the Atlantic slopes of
54 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
the Guyanas (see Vari, 1982a,b for details on those distribu-
tions). The derived characters advanced as synapomorphies for
Curimatopsis by Vari (1982a) are as follows:
35. Expansion of the neural and haemal spines of the two
posterior free preural centra into transversely flattened
plates in the males.
36. Presence of basal spurs on the middle caudal-fin rays in
the males.
37. Expansion of the penultimate principal caudal-fin ray of
the lower lobe of the caudal fin in the males.
38. Ventral reorientation of the ventralmost ray of the upper
caudal fin lobe.
39. Ventral expansion of the metapterygoid and the conse-
quent reduction in the relative size of the metapterygoid-
quadrate fenestra (Figure 42A).
40. Reduction to a single pair of uroneurals.
41. Absence of an ossified antorbital.
42. Reduction in the relative sizes of the supraorbital and the
sixth infraorbital (dermosphenotic).
These synapomorphies are here supplemented by additional
derived characters that also delimit the clade formed by the
five species of Curimatopsis including C. myersi:
43. Oblique angle of articulation between the third infra-
pharyngobranchial and third epibranchial (Figure 15).
44. Very oblique angle of articulation between the second
infrapharyngobranchial and second epibranchial, and the
associated expansion of the second infrapharyngobranchial
posterolaterally (Figure 15).
45. Possession of narrow rather than broad flanges along the
ventrolateral margins of the urohyal.
46. Lack of a laterosensory canal segment in the first
infraorbital.
47. Restructuring of the premaxilla into a very slender
ossification.
48. Enlarged paddle-shaped body of the maxilla (Figure 38A).
49. Reduction of the ethmo-palatine cartilage to a relatively
small body at the tip of the premaxilla (Figure 40A) .
50. Reduction in the degree of scale poring of the laterosen-
sory canal system of the body.
An additional possible synapomorphy for the members of
Curimatopsis is the relatively elongate urohyal extending
distinctly beyond the point of lateral divergence of the
branchiostegal rays (see discussion under "Hyoid Arch").
Vari (1982a: 10) proposed a phylogenetic scheme for the
four species of Curimatopsis recognized in that study. As
mentioned previously, an additional species (C myersi) from
the Rio Paraguay basin was discovered and subsequently
described (Vari, 1982b). The phylogenetic scheme for the
expanded Curimatopsis is presented in Figure 45. Vari (1982a)
discussed the following synapomorphies shared by Curimatop-
sis macrolepis and C. microlepis:
A. Lengthening of the postorbital portion of the head.
B. Terminal, upturned mouth with the lower jaw overlapping
the anterior margin of the upper.
microlepis macrolepis myersi crypticus evelynae
-I-K
-35-50
FIGURE 45.?Qadogram of the most parsimonious hypothesis of relationships within the genus Curimatopsis.
Numbered and lettered synapomorphies correspond to text
NUMBER 471 55
During the course of this analysis those two synapomorphies
have been supplemented by an additional character that was
not discussed by Van (1982a):
C. Expansion of the anteroventral portion of the posttemporal
and the lengthening of the segment of the laterosensory
canal associated with the region of the bone.
The anteroventral region of the posttemporal is relatively
small and has a short laterosensory canal segment in the
Prochilodontidae, other species of Curimatopsis, and other
genera of the Curimatidae. Given the unique nature of the
expansion of that region of the bone in C. microlepis and C.
macrolepis among curimatids, that correlation is hypothesized
to be a synapomorphy for those two species. Similarly the
elongation of the associated laterosensory canal segment in
that region of the posttemporal is also hypothesized as
synapomorphous for those species (SYNAPOMORPHY C).
The clade consisting of Curimatopsis evelynae, C. crypticus,
and C. myersi is defined by three characters previously
proposed by Van (1982a):
D. Pronounced expansion of the penultimate principal ray of
the lower lobe of the caudal fin in males.
E. Possession of a hypural 2 that is separate from the fused
PUj + Uj, but fused to hypural 1.
F. Contact in the males of hypural 3 with the fused hypural
1 and 2, and the development posteriorly of interdigita-
tions between those elements.
Additional synapomorphies for that clade are as follows:
G. Absence of an ossified sixth infraorbital (IOg).
H. Laterally expanded longitudinal flanges on the midlateral
surface of the posterior two free preural centra and the
fused PU, + U,.
Autapomorphies for Curimatopsis evelynae discovered since
Van (1982a) are primarily reductive, and given the small adult
body size of the species may be the consequence of a
paedomorphic process. These characters, which were not
discussed by Vari (1982a), are as follows:
I. Reduction in the degree of ossification of the laterosensory
canals in the second infraorbital.
J. Reduction in the degree of ossification of the laterosensory
canals in the fourth infraorbital.
K. Absence of an ossified suprapreopercle.
Ossified laterosensory canal segments in the second and
fourth infraorbitals and the ossification of the laterosensory
canal above the preopercle as a separate supraperopercle are
generalized for all other curimatids. Synapomorphies I, J, and
K are consequently considered autapomorphous for C.
evelynae.
Autapomorphies for the remaining Curimatopsis species
have not been discovered, nor have characters been discovered
that allow us to advance a hypothesis of sister species
relationships within the trichotomy formed by C. evelynae, C.
crypticus, and C. myersi.
Potamorhina, Curimata, Psectrogaster, Steindachnerina,
Pseudocurimata, Curimatella, and Cyphocharax
These genera form a clade defined by thirteen synapomor-
phies:
51. Medial bony spur extending posteroventrally from the
medial surface of the main body of the fourth epibranchial.
52. Pronounced anterior reorientation of the dorsal process of
the fourth epibranchial beyond the condition in either the
Prochilodontidae or Curimatopsis (Figures 7-11).
53. Anterior bony spur extending forward from the main body
of the fourth epibranchial. That process serves as the area
of attachment for the ligament extending between the third
and fourth epibranchials (Figures 8, 10, 11, 13, 14).
54. Posterior lengthening of the fifth upper pharyngeal tooth
plate, and the reduction of the dentition associated with
that ossification (Figures 6-11).
55. Transversely subdivided posterior articular cartilage on the
third infrapharyngobranchial (Figures 8, 10, 11).
56. Reduction or elimination of the gap between the articular
cartilages on the anterior margin of the second hypobran-
chial (Figure 22B-D).
57. Enlarged ventral process on the fourth ceratobranchial
that serves as a point of attachment for ligamentous tissues
associated with the ventral aorta (present in Potamorhina,
Curimata, and Psectrogaster, secondarily lost (SYNAPO-
MORPHY 93, below) in Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax).
58. Expansion of the portion of the ligamentum primordiale
proximate to the Aj portion of the Adductor Mandibulae
muscle, either with or without an included cartilage body.
59. Opposing articular facets on the premaxilla and maxilla
at their points of contact.
60. Expanded posterodorsal process of the main body of the
maxilla (Figure 38B).
61. Sigmoid dorsal margin of the dentary (Figure 39B).
62. Subdivision of the articular cartilage of the palatine into
an anterior portion contacting the ethmopalatine cartilage
and vomer, and a posterior section with a dorsal articular
surface that contacts the ventral articular process of the
lateral ethmoid (Figure 40B).
63. Band-like ligament between the neurocranium and suspen-
sorium with a discrete area of attachment to both the vomer
and the mesopterygoid.
The large number of synapomorphies for this clade represent
major restructurings of the gill arches, jaws, and associated
soft tissues relative to the state of those systems in
Curimatopsis. These perhaps reflect the increased specializa-
tion for a microphagous diet in the species of Potamorhina,
Curimata, Psectrogaster, Steindachnerina, Pseudocurimata,
56 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
Curimatella, and Cyphocharax. These modifications involve a
number of systems which undergo further modifications at less
inclusive levels of universality within the family. These
independent evolutionary histories lead me to deal with these
synapomorphies as thirteen independent features rather than
attempting to arbitrarily group them as a reduced number of
composite characters.
The lineage defined by characters 51-63 is, in turn, divisible
into two monophyletic groups: one consisting of the species
of the genus Potamorhina, the other formed by the genera
Curimata, Psectrogaster, Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax.
Potamorhina
The genus Potamorhina consists of five relatively large-
bodied species. Potamorhina laticeps is limited to the Lago
Maracaibo basin. P. squamoralevis inhabits the Rio Paraguay
basin. Potamorhina latior and P. pristigaster are endemic to
the Rio Amazonas system. P. altamazonica is common to the
Rio Amazonas system and the Rio Orinoco (see Vari, 1984a,
for further details on species distribution). Potamorhina was
defined by Vari (1984a: 13) on the basis of six synapomorphies,
listed below. Synapomorphy 68 was described in detail by
Vari (1984a) and is not discussed elsewhere in this paper. The
remainder are described herein under "Character Description
and Analysis":
64. Horizontal lengthening of the anterodorsally aligned
dorsal process of the fourth epibranchial and its associated
distal cartilage (Figure 10).
65. Contact of the distal cartilage of the fourth epibranchial
with the cartilage of the uncinate process of the third
epibranchial (Figure 10).
66. Horizontal elongation of the fourth basibranchial, fourth
ceratobranchial and the complex formed by the fourth
infrapharyngobranchial and fourth upper pharyngeal tooth
plate (Figure 10).
67. Elaboration of the laterosensory canal system in the sixth
infraorbital from a tripartite system into one with four or
more branches (Figure 36c).
68. Presence of two or three unbranched rays and 11-16
branched rays in the anal fin (Vari, 1984a:7).
69. Possession of 85-110 pored lateral line scales from the
supracleithrum to the hypural joint.
The additional synapomorphies for the species of Pota-
morhina discovered during this study are as follows:
70. Lengthening of the fifth epibranchial into a long anteriorly
tapering body (Figure 10).
71. Subdivision of the articular process on the uncinate
process of the first epibranchial (Figure 21).
72. Posterior elongation of infraorbitals four and five (IO4 and
IO5) and the associated lengthening of the posterior
horizontal laterosensory canal segment in the fourth
infraorbital.
73. Greater posterior development and more oblique orienta-
tion of the opercle.
74. Posterior development of the interopercle into a large
triangular plate between the preopercle, subopercle, and
opercle.
An additional possible synapomorphy for the members of
Potamorhina is the relatively elongate urohyal extending
distinctly beyond the point of lateral divergence of the
branchiostegal rays (see discussion under "Hyoid Arch").
Vari (1984a: 13) advanced a hypothesis of relationships
among the five recognized species of Potamorhina. Only one
additional character pertinent to that hypothesis was uncovered
during this study and it is congruent with the previously
proposed phylogenetic scheme for and within Potamorhina,
(see Vari, 1984a:4-ll, fig. 6).
The additional derived character discovered during this
study is an autapomorphy for Potamorhina pristigaster. This
feature involves the marked shift posteriorly of the fourth and
fifth infraorbitals and the consequent more acute angle between
the axes of the laterosensory canal segments of those elements.
As discussed previously, this shift is somewhat similar in
overall appearence to that in species of Steindachnerina,
Pseudocurimata, Curimatella, and Cyphocharax (SYNAPOMOR-
PHY 95). The morphological plan of these elements in
Potamorhina pristigaster is a consequence of a posterior shift
in the position of the elements relative to the orbit, and differs
in various details from the condition in the cited genera. In light
of those differences and within the context of the overall most
parsimonious hypothesis of relationships, the state in P.
pristigaster is considered non-homologous with that in
Steindachnerina, Pseudocurimata, Curimatella, and
Cyphocharax.
Curimata, Psectrogaster, Steindachnerina,
Pseudocurimata, Curimatella, and Cyphocharax
The sister lineage to Potamorhina is a clade consisting of
six genera: Curimata, Psectrogaster, Steindachnerina, Pseu-
docurimata, Curimatella, and Cyphocharax. That lineage is
delimited by three shared derived characters:
75. Elimination of the gap between the anterior articular
cartilages along the anterior margin of the second
hypobranchial with the fusion of the cartilage bodies into
a single mass (Figure 22c-D).
76. Development of a buccopharyngeal complex consisting
of three longitudinal fleshy folds on the roof of the oral
cavity, or a further derived elaboration of that complex.
77. Well developed ridge on the medial surface of the
metapterygoid (Figures 41,42).
NUMBER 471 57
The clade delimited by synapomorphies 75-77 contains two
subunits defined by less universal derived characters. These
are Curimata on the one hand, and the clade consisting of
Psectrogaster, Steindachnerina, Curimatella, and Cypocharax
on the other.
Curimata
The genus Curimata is a distinctive group of moderate to
large sized species inhabiting the Rio Magdalena, the Rio
Orinoco, the rivers of the Atlantic slopes of the Guyanas, some
of the rivers of northeastern Brazil, and the Amazon basin.
Curimata is characterized by six synapomorphies:
_? Ontogenetic reconfiguration of the medial spur on the
fourth epibranchial into a ventrally expanded process
contacting the medial margins of the fifth upper pharyn-
geal tooth plate (Figures 7, 8).
79. Development of a median shelf on the dorsal margin of
the spur of the fourth epibranchial, and the associated
pronounced fenestration of that spur (Figure 8).
80. Medial shift of the medial spur and ventral articular
processes of the fourth epibranchial, and of the associated
fifth upper pharyngeal tooth plate (Figure 14).
81. Reduction or loss of the first infrapharyngobranchial.
82. Elaboration of the three primary folds of the bucco-
pharyngeal complex into large ventral flaps, and the
development of numerous secondary parallel folds on the
roof of the buccal cavity (Figure 28).
83. Thickened anteromedial portion of the metapterygoid
(Figure 42D).
An additional possible synapomorphy for the members of
Curimata is the relatively short urohyal extending only to the
point of lateral divergence of the branchiostegal rays. The
morphology of that system may represent a secondary reversal
from the condition in a more inclusive clade (see discussion
under "Hyoid Arch").
Psectrogaster, Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax
These genera, comprising the sister group to Curimata, share
two derived characters:
84. Possession of a discrete cartilage embedded in an
aponeurotic covering within the expanded basal portion
of the ligamentum primordiale (Figure 37).
85. Posteroventral expansion of the mesopterygoid into an
elongate process extending into and partially or totally
subdividing the metapterygoid-quadrate fenestra (Figure
41B).
Synapomorphies 84 and 85 unite two clades: that recognized
herein as Psectrogaster, and the lineage consisting of
Steindachnerina, Curimatella, Pseudocurimata, and
Cyphocharax.
Psectrogaster
The species of Psectrogaster are distributed within the Rio
Orinoco, Essequibo River, Amazon basin, some of the rivers
of northeastern Brazil, and the Rio Paraguay system. Psec-
trogaster species have in common the following seven
synapomorphies:
86. Posterior elongation of the medial spur on the main body
of the fourth epibranchial (Figure 9).
87. Elongation and distal twisting of the fifth upper pharyngeal
tooth plate (Figure 9).
88. Well developed longitudinal ridge on the ventral surface
of the third epibranchial.
89. Well developed longitudinal ridge on the ventral surface
of the second epibranchial.
90. Distinct flange on the dorsal surface of the second
infrapharyngobranchial that overlies the anteromedial
portion of the third infrapharyngobranchial.
91. Distinct longitudinal ridge on the ventral surface of the
first epibranchial.
92. Longitudinal ridge on the dorsal surface of the third
ceratobranchial.
An additional possible synapomorphy for the members of
Psectrogaster is the relatively elongate urohyal extending
distinctly beyond the point of lateral divergence of the
branchiostegal rays (see discussion under "Hyoid Arch").
Steindachnerina, Pseudocurimata, Curimatella,
and Cyphocharax
The lineage consisting of these genera is characterized by
four synapomorphies:
93. Secondary loss of the enlarged ventral process on the
fourth ceratobranchial that is found in Potamorhina,
Curimata, and Psectrogaster (SYNAPOMORPHY 57).
94. Presence of a basihyal tooth plate.
95. Reorientation of the canals in the fourth and fifth
infraorbitals (IO4 and IO5) with the consequent reduction
in the angle of the primary axes between the laterosensory
canal segments in those ossifications (Figure 35c).
96. Reduction of the laterosensory canal system in the sixth
infraorbital (IOg) to a single tube extending between the
fifth infraorbital and pterotic laterosensory canal segments
(Figure 36 D).
An additional character possibly synapomorphous for the
lineage consisting of Steindachnerina, Pseudocurimata, Curi-
matella, and Cyphocharax involves the form of the urohyal.
These genera are characterized by a relatively short urohyal
58 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
that reaches posteriorly to the point of lateral divergence of the
branchiostegal rays. Three alternative, equally parsimonious,
hypotheses exist to explain the occurrence of the derived
condition, the more elongate urohyal, in Curimatopsis,
Potamorhina, and Curimata. One of these alternatives involves
the derived secondary reduction of the urohyal in the clade
consisting of Steindachnerina, Pseudocurimata, Curimatella,
and Cyphocharax (the other alternatives are discussed under
"Hyoid Arch").
It is not possible to resolve the hypothesis of relationships
among the remaining genera in the family {Steindachnerina,
Pseudocurimata, Curimatella, and Cyphocharax) on the basis
of the characters uncovered during this study. Synapomor-
phous characters do, however, permit the definition of three
of these groups {Steindachnerina, Pseudocurimata, and
Curimatella) as individual monophyletic units. The remaining
genus, Cyphocharax, is not delimited by any known derived
characters, and it is thus possible that some or all of its
component taxa may actually be more closely related to
Steindachnerina, Pseudocurimata, or Curimatella, or some
combination of these genera than to the remaining species of
Cyphocharax (see also discussion of Cyphocharax, below).
These four remaining recognized genera are discussed
sequentially, with the order of presentation not indicative of
any underlying hypothesis of relationships.
Steindachnerina
Steindachnerina is the largest genus in the Curimatidae,
with a geographic distribution covering much of the known
range of the family in South America east of the Andes together
with a single species (S. atratoensis) to the west of the Andean
cordilleras in the Rio Atrato system of Colombia. The
hypothesized synapomorphies for the species of Steindachner-
ina are as follows:
97. Ventral and dorsal ridges on the lateral surface of the
second infrapharyngobranchial that bracket the anter-
omedial portion of the third infrapharyngobranchial.
98. Expansion of the cartilaginous portion of the first
infrapharyngobranchial and its contact with the second
infrapharyngobranchial.
99. Attachment of the ligament between the second and third
hypobranchials to a distinct anterior process on the
anterolateral surface of the ventral process of the third
hypobranchial (Figure 24).
100. Lateral expansion of the anterior portion of the basihyal
and associated basihyal tooth plate (Figure 33B).
Pseudocurimata
The species of Pseudocurimata are limited to the trans-
Andean drainages of northwestern South America from the Rio
Atrato of Colombia to the Rio Chira of northern Peru. This
distribution is nearly totally allopatric to the remainder of the
family other than for the partial overlap of the range of P.
lineopunctata with that of Steindachnerina atratoensis in the
Rio Atrato system. Synapomorphies for the genus are as
follows:
101. Expansion of the portion of the second hypobranchial
proximate to the third basibranchial and the fission of the
articular cartilage on the anterior and medial surfaces of
the second hypobranchial (Figure 22D).
102. Posterior shift of the point of interdigitation of the first
proximal pterygiophore of the dorsal fin to between the
fifth and sixth, or sixth and seventh, neural spines
(Figure43B-C).
103. Pronounced reduction or complete loss of the second set
of uroneurals.
Curimatella
The relatively few species (probably six or seven) of
Curimatella occur in the rios Orinoco, Amazonas, Essequibo,
S3o Francisco, and the La Plata basin. The genus is
distinguished by the presence of the following synapomorphy:
104. Layer of scales extending across the lobes of the caudal
fin.
Cyphocharax
Taxa assigned to Cyphocharax in this study are widely
distributed through nearly the entire range of the family
Curimatidae with the exception of the trans-Andean rivers to
the west and south of the Rio Magdalena of Colombia that are
inhabited by species of Pseudocurimata. Cyphocharax species
are also absent in the rivers of northeastern Brazil.
The characters discovered during this study have not
included any synapomorphies congruent with a hypothesis of
the monophyly of Cyphocharax. The combination of that lack
of any defining characters for Cyphocharax and the multicho-
tomy that exists between that genus, Curimatella, Pseudocuri-
mata, and Steindachnerina leaves open the possibility that
some components of Cyphocharax may be more closely related
to one or more of the other genera in the multichotomy than
to the remaining species of Cyphocharax.
Although Cyphocharax is not readily definable on the basis
of derived characters, there are some recognizable assemblages
within that genus that are to varying degrees phyletically and
phenetically distinguishable and may represent monophyletic
subunits of the genus. The largest of these involves eleven
nominal taxa characterized by a distinctive dark midlateral
spot on the caudal peduncle, such as is typical for Cyphocharax
spilurus. Other nominal taxa of the C. spilurus species group
sharing this evidently derived character are C. spiluropsis, C.
gillii, C. surinamensis, C. helleri, C. stigmaturus, C. esperan-
zae, C. ucayalensis, C. vandellii, C. esperanzae pijpersi, and
NUMBER 471 59
C. spilota (see Table 2 for information on authors and original
generic placement of these taxa).
A second assemblage consists of curimatids of relatively
small adult body size, the first described being C. saladensis.
The C. saladensis species group consists of that species, C.
punctata, and C. vanderi all of which share a reduction in the
extent of the poring in the lateral line, a less developed
laterosensory canal system of the head, in particular of the
canal segments in the infraorbitals, and a general reduction in
the degree of development of various cranial sculpturing. These
paedomorphic characters cannot be evaluated here as to the
particular heterochronic process that has resulted in these
results, but are potentially derived characters that unite these
species into a monophyletic group (see Weitzman and Vari,
1988, for further discussion of reductive characters and of their
utility in phylogenetic reconstructions).
Another major assemblage, the Cyphocarex gilberti species
group, includes species sharing overall distinctive pigmenta-
tion patterns, including a rhomboidal caudal pigmentation
pattern in juveniles and random body spotting, which are
possibly derived. The nominal members of the C. gilberti
species group are C. gilberti, C. voga, C. albula, C. grandocule,
C. santacatarinae, and C. modesta (see Table 2 for authors and
original generic assignments). Cyphocharax platana and C.
nagellii, in turn, are phenetically very similar species with
distinctive body forms and having contiguous ranges in the
La Plata and upper Rio Parang systems respectively. These
may be sister species. Groupings among the remaining nominal
Cyphocharax species are not as apparent at this time.
Cyphocharax Fowler (1906) as utilized in this study is
considered to have four synonyms: Xyrocharax Fowler
(1913b), Hemicurimata Myers (1929), Curimatoides Fowler
(1940), and Cruxentina Fernandez-Yepez (1948). Cyphocharax,
Xyrocharax, Hemicurimata, and Curimatoides all have type
species which either represent the same species or very closely
related species, all of which belong to the C. spilurus species
group. Xyrocharax has not been utilized by authors other than
Fowler (1913b), whereas Curimatoides is evidently based on
a single aberrant individual lacking an adipose dorsal fin.
Given the lack of resolution of the phylogeny at this point
and the clumping of the vast majority of the type species of the
available nominal genera in a small subunit of this assemblage
it is most conservative to recognize a single genus for this
questionably monophyletic assemblage. The alternative, the
recognition of all nominal available genera, or an intermediate
number of taxa would necessitate the creation of a series of
new genera for the different species groupings described above
for which generic names have not been previously proposed.
Such an elaboration of the supraspecific taxonomy within this
assemblage is considered inappropriate given the phylogenetic
uncertainties involving interrelationships among the species
united within Cyphocharax at this time, and in light of the
absence of any advantages to such a more complex taxonomic
scheme.
Unassigned
One feature, involving the elongation of the urohyal, was
assigned a synapomorphy number for inclusion in the PAUP
analysis and for discussion purposes (see "Urohyal" and
"Convergent Characters"), but does not appear on the
cladogram (Figure 44). The reason is that three interpretations
of its behavior on the cladogram, each involving three
independent derivations or losses, are equally parsimonious.
The feature is as follows:
105. Elongation of the urohyal past the point of lateral
divergence of the branchiostegal rays in Curimatopsis,
Potamorhina, and Psectrogaster.
Convergent Characters
The hypothesis of phylogenetic relationships within the
family Curimatidae proposed in this study is the most
parsimonious derivable from those characters showing discrete
variability within the family when considered in the context
of the available information on character homology, polarity,
and distribution. The utilization of the principle of simplicity
(parismony), as the criterion to evaluate the preferability of
alternative hypotheses has been the subject of some contro-
versy (e.g., Felsenstein and Sober, 1986). However, this
procedure, the acceptance of the hypothesis that requires the
fewest ad hoc assumptions, is still preferable (Beatty and Fink,
1979; Eldredge and Cracraft, 1980; Wiley, 1981). I recognize,
nonetheless, that evolutionary processes are not necessarily
parsimonious, with the result that evident homoplasies are
features of most phylogenetic schemes including the one
proposed in this study.
The hypothesis of phylogenetic relationships described
herein includes features, from diverse body systems, that have
a phyletic distribution incongruent with the final most
parsimonious phylogenetic scheme. Such incongruent derived
characters, homoplasies, will be evaluated here within two
contexts. The first framework deals solely with the phylo-
genetic distribution of the characters regardless of the type of
character involved. Do the homoplasies under consideration
characterize two or more independent sublineages within the
Curimatidae (convergences internal to the Curimatidae), or do
they represent the independent acquisition of the character in
question both in a subunit of the Curimatidae and in some
characiform outgroup (convergences external to the Curimat-
idae)? Those comparisons serve to emphasize the phyletic
patterns of homoplasy and focus attention on those clades that
evidently have independently undergone congruent patterns
of convergent evolution. Admittedly the distinction between
internal and external homoplasies is based on arbitrary
taxonomic limits, but serves nonetheless to convey a general
sense of the phyletic proximity of the clades involved in the
comparisons.
One can alternatively discuss homoplasies in terms of their
60 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
innovative or reductive natures (Weitzman and Fink, 1985)
outside of a strictly phyletic context. Innovative characters,
having "additional features beyond those present in the
outgroups" (Weitzman and Fink, 1985:10), are a consequence
of changes in peramorphic processes and most commonly
represent terminal additions in overall developmental se-
quences. Reductive characters in contradistinction involve the
reduction or loss of a feature present in the outgroups and may
result from a number of different paedomorphic developmental
changes resulting from progenesis, neoteny, and post-
displacement. The analysis of common patterns of reductive
homoplasies within a phylogentic scheme derived from
multiple characters, innovative and reductive, congruent and
incongruent, permits an evaluation of the utility of at least
some reductive characters as indicators of phylogenetic
relationship.
Since the homoplasies to be discussed within these two
contexts (phyletic, and innovative versus reductive) are the
same, they will only be detailed in the first set of discussions,
those on phyletic distribution. Homoplasies in that discussion
are divided between internal homoplasies and external
homoplasies, and are listed sequentially by synapomorphy
number. These synapomorphies will also be discussed in the
context of reductive versus innovative homoplasies.
The homoplasies found in two or more subunits of the
Curimatidae (internal homoplasies) and their phylogenetic
distributions are as follows, listed by synapomorphy number
(see also "Character Description and Analysis" and "Synapo-
morphy List and Phylogenetic Reconstruction"):
21. Subdivided articular processes on the uncinate process
of the first epibranchial in Potamorhina and the lineage
consisting of Curimata incompta, C. cyprinoides, C.
copei, C. schomburgkii, C. edentulus, C. planirostris,
C. kneri, C. mivartii, C. cerasina, C. aspera, C. simulata,
C. cisandina, C. alleni, and two undescribed species of
Curimata.
28. Elaboration of the buccopharyngeal complex in Curi-
mata species and the majority of Steindachnerina
species.
46. Absence of a laterosensory canal segment in the first
infraorbital in all Curimatopsis species, Steindachnerina
binotata, Cyphocharax gillii, C. saladensis, C. punctata,
and C. vanderi. This reduction also occurs in various
characiform outgroups but not in the proximate out-
groups Prochilodontidae, Anostomidae, and Chilodontidae.
50. Reduction in the degree of poring of the laterosensory
canal system on the body in Curimatopsis and in three
species of Cyphocharax {vanderi, punctata, and saladen-
sis); also reduced in some other characiforms, but not in
the proximate outgroups to the Curimatidae.
69. High number of lateral-line scales in Cyphocharax
abramoides and in the species of Potamorhina.
88. Well developed ventral longitudinal ridges along the
ventral surface of the third epibranchial in Curimata
aspera, C. simulata, C. cerasina, and all Psectrogaster
species.
89. Well developed ventral longitudinal ridges along the
ventral surface of the second epibranchial in Curimata
mivartii, C. aspera, C. simulata, C. cerasina, C.
cisandina, C. alleni, two undescribed Curimata species,
and all species of Psectrogaster.
90. Distinct flange on the anterodorsal surface of the second
infrapharyngobranchial in Psectrogaster (see also dis-
cussion of Synapomorphy 97 and under "Third Epibran-
chial").
91. Longitudinal ridge along the ventral surface of the first
epibranchial in Psectrogaster and in two species of
Curimata.
92. Longitudinal ridge along the dorsal surface of the third
ceratobranchial in Psectrogaster and in Curimata mivar-
tii, C. aspera, C. simulata, C. cerasina, C. alleni, and
two undescribed species of Curimata.
94. Presence of a basihyal tooth plate in a subset of species
of Curimata and in the vast majority of the species in the
lineage consisting of Steindachnerina, Pseudocurimata,
Curimatella, and Cyphocharax.
97. Distinct flange on the anterodorsal surface of the second
infrapharyngobranchial in Steindachnerina (see also 91,
above, and discussion under 'Third Epibranchial").
102. Posterior position of the interdigitation of the first
proximal pterygiophore of the dorsal fin with the
proximate neural spines in Curimata ocellata, C.
semitaenitata, and all Pseudocurimata species (also
present in some outgroups, see "Supraneurals").
105. Elongation of the urohyal past the point of lateral
divergence of the branchiostegal rays in Curimatopsis,
Potamorhina, and Psectrogaster. Three alternative, equally
parsimonious hypothesis exist (see discussion under
'Urohyal'). The first is that a lengthened urohyal is
synapomorphous for all species of the Curimatidae and
is secondarily independently shortened in Curimata and
the lineage consisting of Steindachnerina, Pseudocuri-
mata, Curimatella, and Cyphocharax. The elongation
may also be acquired at the level of the Curimatidae,
with a reduction in the clade consisting of Curimata,
Psectrogaster, Steindachnerina, Pseudocurimata, Curi-
matella, and Cyphocharax, with a secondary lengthening
in Psectrogaster. Finally, the elongation may have been
independently acquired within each of the three lineages,
Curimatopsis, Potamorhina, and Psectrogaster. The
feature was not used in the cladogram because of the
several possible interpretations of its distribution.
As can be seen in the above listing, a number of generic and
subgeneric subunits possess characters homoplasious relative
to the overall most parsimonious hypothesis of intrafamilial
relationships. Nonetheless, relatively few patterns of repeated
NUMBER 471 61
homoplasy among those taxa were discovered during the
course of this study. Reductive features of the laterosensory
canal system of the head and body are common to
Curimatopsis species and a subunit of Cyphocharax (SYNAPO-
MORPHIES 46 and 50, see also discussion in next section).
Psectrogaster and subunits of Curimata of varying levels of
inclusiveness also share several convergences involving
longitudinal ridges on various gill arch elements (SYNAPOMOR-
PHIES 89,92 at one level, and 89, 91 at another). The differing
levels of inclusiveness at which these convergences are
common in those genera, along with overall parsimony
considerations indicate that these are homoplasious rather than
homologous features.
The second subset of homoplasies are derived synapomor-
phies occurring in a subunit of the Curimatidae and also in a
characiform outgroup (external homoplasies). These are as
follows:
16. Absence of jaw teeth in all members of the Curimatidae,
and in Anodus and Eigenmannina of the Hemiodontidae.
39. Reduction in the size of the metapterygoid-quadrate
fenestra as a consequence of the ventral expansion of the
metapterygoid in Curimatopsis, and in the Chilodon-
tidae.
41. Absence of an ossified antorbital in Curimatopsis in the
Curimatidae, and in the African characid Lepidarchus
adonis.
100. Expansion of the basihyal and basihyal tooth plate in
Steindachnerina, and in the Prochilodontidae.
104. Sheet of scales across the lobes of the caudal fin in
Curimatella, and in the genus Leporellus in the
Anostomidae.
Synapomorphies 46, 50, and 102, homoplasious within the
Curimatidae, are also synapomorphous for some lineages
outside of that family. Only synapomorphy 102 is homopla-
siously present in proximate outgroups of the Curimatidae.
The above listing of external homoplasies should not be
considered exhaustive since it is not feasible to carry out
outgroup comparisons against the entire spectrum of characi-
form outgroups. These comparisons have for practicality been
focused on the proximate sister groups to the Curimatidae, the
Prochilodontidae, Anostomidae, and Chilodontidae.
An alternative way to look at the above attributes is to ask
whether they are innovative or reductive, and whether these
classes of homoplasies demonstrate a meaningful pattern
relative to the arrived at phylogenetic hypothesis. The
innovative characters in the previous listings having a
homoplasious distribution internal to the Curimatidae are
numbers 28, 69, 88-92, 94, 102, and 105. The innovative
characters in the previous listing having a homoplasious
distribution external to the Curimatidae are 39, 100, and 104.
The innovative homoplasies are common to some members
of the Curimatidae and a diversity of outgroups. These do not,
however, demonstrate any patterns that would support
alternative hypotheses of relationships. This lack of congruence
makes it unfruitful to speculate on common underlying factors
resulting in these homoplasies.
The reductive homoplasies, internal to the Curimatidae,
uncovered during this study are synapomorphies 46 and 50.
Those external to the Curimatidae are synapomorphies 16 and
41.
Reductive characters and their utility in phylogenetic
reconstructions in fishes in general and characiforms in
particular have been the subject of discussions by various
authors (e.g., Weitzman and Fink, 1983; Weitzman and Vari,
1988). As noted by those authors, paedomorphic features
cannot be a priori distinguished from primitive states in the
absence of information on phylogenetic relationships derived
from multiple character systems involving innovative and
reductive apomorphies. The bulk of the characters used by
different authors in advancing hypotheses of relationships in
characiforms involve external characters. In the Curimatidae
the external reductive features most often used for those
purposes have involved the extent of development of the
laterosensory canal system, in particular the degree of scale
poring associated with the lateral line on the body.
The use of a reduction in the degree of poring on the lateral
line to distinguish a species at the supraspecific level within
the Curimatidae was first advanced by Steindachner (1876)
who proposed Curimatopsis as a subgenus of Curimata in his
description of C. macrolepis, a species that has scale poring
developed only on the anterior portion of the laterosensory
system of the body. That character was, in turn, used by
Eigenmann and Eigenmann (1889b) as part of their definition
of Curimatopsis as a genus and in their description of
Curimatopsis microlepis, by Ahl (1931) when proposing
Curimatopsis macrocephalus, and by G6ry (1964a) for his
Curimatopsis evelynae. Other authors who have also used
Curimatopsis for species with incompletely pored lateral lines
are Meinken (1933) in his description of Curimatopsis
saladensis and Ahl (1934) when he described Curimatopsis
maculatus. Vari (1982a) redefined Curimatopsis on the basis
of a number of derived innovative characters and described C.
crypticus, a species with an incompletely pored lateral line.
That attribute also characterized C. myersi, subsequently
described by Vari (1982b). Both Curimata vanderi Britski
(1980) and C. punctata Vari and Nijssen (1986) have lateral
line scale poring limited to the anterior portion of the body,
but were not placed by their authors in Curimatopsis.
The nominal species with incomplete poring of the
laterosensory canal system of the body fall into two distinct
groups, those assignable to Curimatopsis as defined by Vari
(1982a) and in this study (C. macrolepis, C. microlepis, C.
microcephalus ( = C. macrolepis), C. evelynae, C. crypticus,
and C. myersi), and those assigned herein to Cyphocharax (C.
saladensis, C. maculatus, C. vanderi, and C. punctata). The
62 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
Curimatopsis species are definable by a number of unique
innovative synapomorphies (see "Synapomorphy List and
Phylogenetic Reconstruction"). Cyphocharax, although not
itself delimited by any unique characters, does share numerous
synapomorphies at different levels of universality with
Potamorhina, Curimata, Psectrogaster, Steindachnerina, Cu-
rimatella, and Pseudocurimata; characters not found in
Curimatopsis species. Thus the data overwhelmingly indicate
that the clade recognized as Curimatopsis and the assemblage
of species of Cyphocharax with incomplete poring of the lateral
line are not closely related to each other (Curimatopsis
maculatus Ahl, which has an incomplete lateral line, is based
on a juvenile of a Cyphocharax species that has a fully pored
lateral line as an adult). The reduced development of portions
of the laterosensory canal system of the body in the two
assemblages is considered homoplasious, although constituting
a synapomorphy for each of the assemblages at lower levels
of universality. A similar situation involves the absence of the
laterosensory canal segment in the first infraorbital; a reduction
that characterizes the species having the reduced poring of the
lateral line and in addition Steindachnerina binotata. Again the
absence of that canal segment is judged to be homoplasious
within the family given the overall most parsimonious
hypothesis of relationships within the Curimatidae. That
character is nonetheless an evident synapomorphy for Curima-
topsis on the one hand, and the assemblage formed by
Cyphocharax saladensis, C. vanderi, and C. punctata on the
other, and is an autapomorphy for Steindachnerina binotata.
Reductive characters present problems relative to their utility
in the generation of phylogenetic hypotheses. Some authors
(e.g., Hecht and Edwards, 1977) have noted the potential
difficulties in determining homologies when losses or reduc-
tions are being studied since it is not possible to critically
compare a character that is absent. Similarly Weitzman and
Vari (1988) discussed the difficulties in discriminating
paedomorphic features from primitive characters in the absence
of a phylogenetic framework. Those problems are well
exemplified by the reductions in the laterosensory canal
segments noted above, and the need for phylogenetic studies
to evaluate such patterns emphasized by Weitzman and Vari
is well demonstrated. By focusing on a single reductive body
system, the reduced lateral line poring, Meinken (1933) and
Ahl (1934) aligned their nominal species (C. saladensis and
C. maculata respectively) with Curimatopsis (sensu stricto).
The phylogenetic hypothesis based on a more extensive set of
characters, internal and external, indicates in contradistinction
that those reductions were independently achieved in those two
species and Curimatopsis.
Weitzman and Vari (1988) also discussed the correlation
between paedomorphic features and reductions in relative body
sizes. Although none of the curimatid species with reductive
characters fits the size limit for miniatures proposed in that
study, those authors noted that miniaturization must be dealt
with as a relative concept. Miniatures must be judged relative
to the body sizes of their immediate outgroups. Within that
definition, some of the curimatids with reductive features could
perhaps be considered miniatures. Four of the Curimatopsis
species (macrolepis, evelynae, myersi, and crypticus) are all
among the smallest species of curimatids, being much smaller
than any species of the Prochilodontidae, which is the sister
group to the Curimatidae. Species of Curimatopsis are also of
smaller maximum body size than the species of their sequential
proximate sister groups within the Curimatidae: Potamorhina,
whose species achieve 205-265 mm SL; Curimata, whose
members reach 112-235 mm SL; and Psectrogaster, the
species of which grow to 130-210 mm SL.
Curimatopsis is definable on the basis of sixteen
synapomorphies. Five of these (SYNAPOMORPHIES 40,41,42,
46, and 50) are reductive, two involving losses of autogenous
ossifications (the second set of uroneurals, SYNAPOMORPHY
40, and the antorbital, SYNAPOMORPHY 41). The other three
reductive synapomorphies for the genus involve reductions in
the degree of development of the laterosensory canal systems
of the head and body, along with a reduction or loss of some
of the dermal ossifications associated with that sensory system.
Although Curimatopsis species do not as adults approach the
standard lengths of members of the Prochilodontidae or
proximate clades to that genus within the Curimatidae, there
is nonetheless more than a two-fold range between the standard
lengths of the largest examined individuals of C. evelynae (40.1
mm SL) and C. microlepis (99.4 mm SL). Across the range of
lengths within the genus there is a correlation between body
size and the degree of expression of various reductive,
apparently paedomorphic features. Curimatopsis microlepis
and C. macrolepis which achieve standard lengths of 99.4
mm and 60.6 mm respectively do not demonstrate any
reductive features beyond those general for the genus. The
species of the sister group to the clade formed by those genera,
the lineage consisting of C. evelynae, C. crypticus, and C.
myersi, range in size from 40.1 mm to 48.7 mm SL. That clade
is characterized in part by an additional reductive feature, the
absence of an ossified sixth infraorbital (Curimatopis SYNAPO-
MORPHY G). The trend to smaller adult body sizes is most
pronounced in Curimatopsis evelynae, which also autapomor-
phously possesses a further series of reductive characters
involving the laterosensory canal system (Curimatopsis
SYNAPOMORPHIES I, J, H). These losses are among those
"typical" for miniatures according to Weitzman and Vari
(1988), and the progressive increase in the number of
paedomorphic features would be expected if there is a
correlation between a reduction in body size and a truncation
of developmental sequences. Further studies are necessary to
determine whether the evident correlation between reduced
size and an increase in apparently paedomorphic features that
is found in Curimatopsis is a general pattern across the
Characiformes.
The concept of miniatures might also be applied to at least
some of the Cyphocharax species with reductive features.
NUMBER 471 63
Again these are species smaller than the other members of the
genus and proximate outgroups, although one species,
Cyphocharax saladensis, does achieve relatively large body
sizes (58.3 mm SL). A more definitive statement on possible
paedomorphosis in Cyphocharax must await the proposal of a
phylogenetic hypothesis for those members of the genus of
reduced body size.
Comparisons with Previous Classifications
A variety of alternative classificatory schemes have been
proposed for, and applied to, the species of the Curimatidae.
One extreme in this century was advanced by Eigenmann
(1910:420-422) who recognized only seven genera for the
known species of Curimatidae, exclusive of Anodus and
Eigenmannina. At the opposite extreme, less than four decades
later Fernandez-Y6pez (1948), describing 17 new genera,
recognized 29 genera. Fowler (1906, 1913b, 1940, 1941)
independently advanced a number of genera that have been
inconsistently recognized since their proposal.
The complexity of the alternative previous classifications
of the Curimatidae makes it inefficient to discuss comparative
merits and problems of each of those systems. Rather I will
only discuss those two arrangements that have been most
widely utilized: Eigenmann (1910), and Fernandez-Yepez
(1948). Independently and in common, these demonstrate
problems typical of the diverse classificatory schemes pro-
posed for the Curimatidae.
Eigenmann's (1910) classification consisted of the follow-
ing: one large genus, Curimatus, containing about 71% of the
66 species then recognized in the family; four genera
{Psectrogaster, Curimatella, Curimatopsis, and Semitapicis)
with two to eight nominal species; and two monotypic genera
(Potamorhina and Gasterotomus). Among the genera recog-
nized by Eigenmann (1910) with two or more species, only
one, Curimatopsis, reflected a monophyletic unit in the
phylogeny proposed in the present study. Psectrogaster as
defined herein was divided by Eigenmann (1910) between that
genus (sensu stricto) and Curimatus (C isognathus and C.
rutiloides). Similarly Curimatella of Eigenmann (1910) agrees
relatively well although not completely with the concept of
that genus proposed in this study, but Curimatus dorsalis
Eigenmann and Eigenmann was retained in Curimatus by
Eigenmann (1910), whereas currently available data indicate
that its closest relatives are the Curimatella species of
Eigenmann's (1910) classification.
Eigenmann's (1910) Curimatus, his most inclusive genus,
has components that are now assignable to nearly all the
curimatid lineages described herein, including the following
genera (with Eigenmann's, 1910, species): Cyphocharax
{Curimatus spilurus, C. gillii)', Curimatella {Curimatus dor-
salis, C. leucostictus); Steindachnerina {Curimatus nasus, C.
conspersa); Pseudocurimata {Curimatus troschelii, C. brevi-
pes); Curimata {Curimatus aspera, C. mivartii); and Psec-
trogaster {Curimatus rutiloides, C. isognathus). Reference to
Figure 44 shows that Curimatus in the sense of Eigenmann
(1910) is thus scattered across the phylogeny proposed here,
and so does not form a monophyletic assemblage. Also,
components placed by Eigenmann (1910) in other nominal
genera (e.g., Curimatella (sensu stricto), Psectrogaster (sensu
stricto), Semitapicis (in part)) would have to be returned to
Curimatus as defined by Eigenmann in order for that genus to
be monophyletic.
This situation exemplifies one of the most common
consequence of classifications predicated primarily on con-
cepts of external differences. The taxa with some externally
obvious, often derived, characters were separated in their own
genera. As a result the remaining species were left in an
assemblage definable only by the absence of the distinctive
features of the excluded taxa. Such taxa consisting of residual
species were in most instances non-monophyletic, or if
monophyletic, could not be recognized as such from data then
available.
Fernandez-Yepez (1948) drastically subdivided the Curimat-
idae into 27 genera (not including Anodus and Eigenmannina)
of which ten were monotypic. That high percentage of
monotypy and the large number of genera with small numbers
of species would make generic level comparisons between that
system and the results of this study tedious and confusing. As
a consequence the best framework to evaluate Fernandez-
Yepez's (1948) classification is to compare the distribution of
species within his classification with the groupings arrived at
in the present study.
No lineages were found to be common to the two schemes
of relationship, with drastic differences between them occur-
ring in most instances. A case by case comparison would again
be tedious but several examples will be given to provide a sense
of the differences. The genus Potamorhina of the present study,
a lineage well corroborated by multiple synapomorphies, is
subdivided between the tribes Potamorhini and Curimatini
under Fernandez-Y6pez's (1948) system. Both of those taxa,
in turn, contained species assigned to all of the other lineages
recognized herein, with the exception of Curimatopsis.
Curimata, as defined herein, was scattered across eight genera
in Fernandez-Ydpez's scheme, with individual Curimata
species aligned by that author with species considered to
belong to Cyphocharax, Steindachnerina, Psectrogaster, and
Curimatella in this paper. Furthermore the individual lineages
containing species of Curimata did not form a discrete subunit
in Fernandez-Yepez's (1948) phylogenetic scheme. Compara-
ble problems with Fernandez-YSpez's (1948) "phylogenetic
tree" exist relative to each of the other lineages recognized in
this study. Thus the congruence of Fernandez-Yepez's (1948)
classification with the phylogeny proposed herein is judged to
be minimal.
To a large extent the cited problems appear to be the result
of Fernandez-Y6pez's (1948) dependence on a few characters
as the basis for his classification. One of the most significant
64 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
of those features was the degree of development of various
prepelvic, postpelvic, and predorsal angles in the body walls,
with that being supplemented by characters involving scalation
on the caudal fin rays, relative fin position, mouth form and
position, and degree of development of the laterosensory canal
system on the body. Many of those noted characters do
constitute synapomorphies at some levels of universality and
thus are potentially useful in generating phylogenetic hypothe-
ses. Nonetheless they also demonstrate a notable degree of
convergence within the Curimatidae in the context of the
phytogeny proposed herein. Such convergence can only be
detected by the examination of a broader spectrum of internal
and external characters than those utilized by Fernandez-Yepez
(1948) and other authors. This problem is best exemplified by
the convergencies in the laterosensory canal system reductions
that have been discussed in the previous sections. Above and
beyond questions of character homology and homoplasy,
Fernandez-Yepez's (1948) classification also demonstrates the
weaknesses already commented on relative to the Eigenmann
(1910) scheme ? the separation of phenetically externally
distinctive taxa into their own monotypic genera, leaving
non-monophyletic residual amalgams of externally generalized
species typically recognized as genera.
Although both authors used primarily external features as
the basis for their classifications, the noted problems in their
taxonomic schemes are not a consequence of the reduced utility
of external characters in phylogenetic reconstructions, but
rather an outgrowth of the methodologies utilized. The
problems noted above for the classifications of Eigenmann
(1910) and Fernandez-Yepez (1948) are, in turn, typical of the
classifications proposed by other authors who have also dealt
with the family (e.g., Fowler, 1975).
Re sumo
A famflia Curim?tidae consiste de mais de 120 especies nominais que habitam uma ampla
seYie de ecossistemas de agua doce neotropicais, variando desde os lagos e rios de curso lento
das terras baixas ate os riachos de curso mais rapido das elevacoes medias ao longo do sope
dos Andes e das terras altas do escudo das Guianas e do escudo do Brasil. Curimatideos ocorrem
a oeste e ao norte dos Andes nos rios das vertentes do Pacifico, desde o sul da Costa Rica ate
a norte do Peru, e nas drenagens do Caribe no noroeste da America do Sul, incluindo os sistemas
do rio Atrato, no Magdalena e lago Maracaibo. Os maiores numeros de especies de curimatideos
ocorrem nas vertentes atlanticas da America do Sul, desde o rio Orinoco, passando pela bacia
Amazonica e os rios costeiros das Guianas, Brasil, Uruguai e Argentina, ate ao sul de Buenos
Aires.
Classificacoes previas da famflia nao foram baseadas em hip6teses de relacoes evolutivas
dentro da famflia. Em lugar disso estas propuseram novos generos e subgeneros com base em
graus de similaridade ou diferenca e, assim mesmo, usando somente caracteres externos. A
"arvore filogenetica" formulada por Fernandez-Yepez (1948) e uma representacao direta de sua
chave dicotomica para os generos da famflia e nao de um verdadeiro esquema filogenetico.
Como consequencia desses e outros problemas, muitas das unidades taxonomicas situadas acima
do nivel de especie reconhecidas em classificoes preVias nao representam agrupamentos
naturais. Os estudos relatados neste trabalho foram realizados tendo em vista a formulacao de
uma hipotese, baseada em caracteres derivados distintos, quanto as relacSes filogeneticas a
nivel supra-especifico dentro da famflia Curimatidae.
O esqueleto e outros sistemas anatomicos das especies da famflia Curimatidae foram
estudados com o fim de se investigar a hip6tese de que esta 6 uma subunidade da ordem
Characiformes, e tambem de se propor uma hipotese quanto as relacoes filogeneticas dentro
da famflia. Sinapomorfias associadas com o aparelho branquial, complexo buco-faringeano,
arco hi6ide, antorbital, infra-orbitais, aparelho opercular, "ligamentum primordiale," maxilas,
arco palatino e neurocranio sao congruentes com a hipotese de que a famflia e uma linhagem
monofil6tica. Caracteres adicionais num subgrupo desses sistemas sao congruentes com a
hipdtese proposta por Van (1983) quanto a Curimatidae e Prochilodontidae serem
grupos-irmaos. A posse em comun de caracteres derivados nos sistemas corporais citados
anteriormente, assim como no complexo hipural, nos raios da nadadeira caudal e na coluna
vertebral definem subunidades naturais dos Curimatidae. Esses caracteres apoiam uma hip6tese
filogen6tica com quatro dicotomias sequenciais e uma politomia terminal. Sete linhagens,
reconhecidas como generos (Curimatopsis, Potamorhina, Curimata, Psectrogaster, Steindach-
nerina, Pseudocurimata, e Curimatella), sao consideradas monofileticas com base na posse
em comum de caracteres derivados. Nao foram encontrados caracteres derivados unicos que
definam Cyphocharax, o oitavo genero reconhecido.
A hipotese para relacoes filogeneticas intra-familiares obtida neste estudo e incongruente,
em diferentes graus, com os esquemas de classificacao propostos por Eigenmann & Eigenmann
(1889b), Eigenmann (1910), Fowler (1975), e Gery (1977b). Existem numerosas diferencas,
em todos os niveis taxonomicos, entre o esquema de interrelacoes proposto neste estudo e a
"arvore filogenetica" apresentada por Fernandez-Yepez (1948). A transferencia, proposta pela
primeira vez por Roberts (1974), dos generos Anodus Spix e Eigenmannina Fowler,
tradicionalmente incluidos nos Curimatidae, para os Hemiodontidae, 6 uma mudanca
congruente com os resultados obtidos por Vari (1983), estando de acordo com a distribuicao
de varios caracteres adicionais considerados como derivados descobertos durante o presente
estudo.
Oito generos: Curimatopsis Steindachner (1876), Potamorhina Cope (1878), Curimata Bosc
(1817), Psectrogaster Eigenmann & Eigenmann (1889a), Steindachnerina Fowler (1906),
Pseudocurimata Fernandez-Yepez (1948), Curimatella Eigenmann & Eigenmann (1889b) e
Cyphocharax Fowler (1906) sao reconhecidos na famflia. Curimatichthys Fernandez-Yepez
(1948) 6 considerado um sinonimo de Curimatopsis. Potamorhina 6 redefinida de maneira a
incluir Gasterotomus Eigenmann (1910), Gasterostomus Femandez-Yepez (1948), Suprasinele-
pichthys Fernandez-Y6pez (1948), e Potamorrhina Braga & Azpelicueta (1983). Curimata
Bosc (1817) tern nove sinonimos juniores: Semitapicis Eigenmann & Eigenmann (1889b),
Peltapleura Fowler (1906), Acuticurimata Fowler (1941), Allenina Fernandez-Yepez (1948),
Lambepiedra Fernandez-Yepez (1948), Bitricarinatra Fernandez-Yepez (1948), Bondichthys
Whitley (1953), Stupens Whitley (1954), e Semitapiscis Braga and Azpelicueta (1983). Tres
generos sao incluidos em Psectrogaster. Estes sao: Pseudopsectrogaster Fernandez-Yepez
65
66 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
(1948), Hamatichthys Fernandez-Yepez (1948) e Semelcarinata FernSndez-Yepez (1948).
Rivasella e Curimatorbis, ambos propostos por Fernandez-Yepez (1948), sao colocados como
sin6nimos de Steindachnerina Fowler (1906). Curimatella tern tres sinonimos juniores:
Apolinarella, Walbaunina e Lepipinna, todos de autoria de Fernandez-Y?pez (1948).
Cyphocharax tem quatro sinonimos: Xyrocharax Fowler (1913b), Hemicurimata Myers (1929),
Curimatoides Fowler (1940) e Cruxentina Fernandez-Yepez (1948).
Caracteres homoplasicos em relacao a filogenia proposta sao discutidos em termos da sua
distribuicao dentro e fora da famflia Curimatidae e tamb6m quanto a serem caracteres
inovadores ou redutivos.
Caracteres redutivos, envolvendo o grau de desenvolvimento do sistema de canais
latero-sensoriais cefalicos e corporais presentes nas especies de Curimatopsis e num grupo de
especies de Cyphocharax, revelaran-se convergentes. Estes caracteres pedomorficos sao
evidentemente uma consequencia da miniaturizacao relativa dessas esp&ies em comparacao
com seus extra-grupos mais pr6ximos.
Literature Cited
Ahl, E.
1931. Neue Siisswasserfische aus dem Stromgebiet des Amazonenstromes.
Sitzungerichte der Gesellschaft Naturforschende Freunde, 1:206-
211.
1934. Beschreibungen zweier neuer Siisswasserfische aus Siidamerika.
Sitzungberichte der Gesellschaft Naturforschende Freunde, 24:
238-241.
Alexander, R. McN.
1964. Adaptation in the Skulls and Cranial Muscles of South American
Characinoid Fish. Journal of the Linnean Society (Zoology),
45(305):169-190.
Allis, E. P., Jr.
1905. The Latero-Sensory Canals and Related Bones in Fishes. Interna-
tional Monatsschrift fur Anatomie und Physiologie, 21:401-463,
figures 1-45.
Azevedo, P. de, M. Vianna Dias, and B. Borgas Vieria
1938. Biologia do Saguiru (Characidae, Curimattnae). Memorias do
Instituto Oswaldo Cruz, 33:481-553, figures 1-12, plates 1-3.
Beatty, J., and W.L. Fink
1979. [Review of] Simplicity, by E. Sober, 1975, Oxford: Clarendon Press.
Systematic Zoology, 28:643-651.
Bertmar, G., B.G. Kapoor, and R.V. Miller
1969. Epibranchial Organs in Lower Teleostean Fishes?An Example of
Structural Adaptation. International Review of General and
Experimental Zoology, 4:1-48.
Bloch, M.E.
1785-1795. Naturgeschichte der Ausldndischen Fische. 324 plates. Berlin.
Bohlke, J.
1953. A Minute New Herring-like Characid Fish Genus Adapted for
Plankton Feeding, from the Rio Negro. Stanford Ichthyological
Bulletin, 5(l):168-170, figure 1.
1958. Studies on Fishes of the Family Characidae, No. 14: A Report of
Several Extensive Collections from Ecuador. Proceedings of the
Academy of Natural Sciences of Philadelphia, 110:1-121.
Bosc, L.A.G.
1817. Nouveau dictionnaire d'Histoire naturelle. 560 pages. Paris.
Boulenger, G.A.
1887. An Account of the Fishes Collected by Mr. C. Buckley in Eastern
Ecuador. Proceedings of the Zoological Society of London,
2:274-283.
1902. Descriptions of New Fishes and Reptiles Discovered by Dr. F.
Silvestri in South America. Annals and Magazine of Natural
History, series 7, 9:284-288.
1904. Fishes (Systematic Account of the Teleostei). In S.F. Harmer and
A.E. Shipley, editors, The Cambridge Natural History, Volume
7:475-727. London: Macmillan and Co. Ltd.
1911. Description of Three New Characinid Fishes from Southwestern
Colombia. Annals and Magazine of Natural History, series 8,
7:212-213.
Braga, L., and M. Azpelicueta
1983. Semitapicis squamoralevis sp. nov. (Osteichthyes: Curimatidae) con
consideraciones sobre el genero. Studies on Neotropical Fauna and
Flora, 18:139-150.
1987. Curimata biornata, A New Curimatid Fish (Characiformes,
Curimatidae) from Argentine and Southeastern Brazil. Revue Suisse
de Zoologie, 94(2):465-473.
Brewster, B.
1986. A Review of the Genus Hydrocynus Cuvier, 1819 (Teleostei:
Characiformes). Bulletin of the British Museum (Natural History),
Zoology, 50(3): 163-206, figures 1-31.
Britski, H.A.
1969. Lista dos Tipos de Peixes dos Colecoes do Departemento de
Zoologia da Secretaria da Agricultura da Sao Paulo. Papeis Avulsos
de Zoologia, Sao Paulo, 22:197-215.
1980. Sobre uma nova especie de Curimata da Bacia do Parana, no Estado
de Sao Paulo (Pisces, Curimatidae). Papeis Avulsos de Zoologia,
Sao Paulo, 33:327-333.
Bussing, W.A.
1966. New Species and New Records of Costa Rican Freshwater Fishes
with a Tentative List of Species. Revista de Biologia Tropical,
14:205-249.
Carvalho. F.M.
1984. Aspectos biologicos e ecofisiologicos de Curimata (Polamorhina)
pristigaster, um Characoidei Neotropico. Amazoniana, 8(4):525-
539.
Cope, E.D.
1878. Synopsis of the Fishes of the Peruvian Amazon Obtained by
Professor Orton During his Expeditions of 1873 and 1877.
Proceedings of the American Philosophical Society, 17:673-701.
Cuvier, G., and A. Valenciennes
1849. Histoire naturelle des poissons. Volume 22, 395 pages, 17 plates.
Paris.
Daget, J.
1960. Le genre Xenocharax (Poissons, Characiformes). Revue de Zoologie
et de Botanique Africaines, 61:35-48, figures 1-10.
1961. Note sur les Nannocharax (Poissons, Characiformes) de l'Ouest
Africain. Bulletin d'Institut Franqais de I'Afrique Noire, 23:165-
181, figures 1-10.
1962a. Le genre Citharinus. Revue de Zoologie et de Botanique Africanines,
66:81-106, figures 1-12.
1962b. Le genre Citharidium. Bulletin d'Institut Franqais de I'Afrique
Noire, Series A, 24(2):505-522.
1964. Le Crane des Teleosteens. Memoires du Museum National
d'Histoire Naturelle, 31(2):163-342, figures 1-53.
Dahl, G.
1971. Los Peces del norte de Colombia. 391 pages. Bogota: Inderena.
Eigenmann, C.H.
1910. Catalogue of the Freshwater Fishes of Tropical and South Temperate
America. Report of the Princeton University Expedition to
Patagonia, 1896-1899, 3(4):375-5U.
1912a. The Freshwater Fishes of British Guiana, Including a Study of the
Ecological Grouping of Species and the Relation of the Fauna of the
Plateau to That of the Lowlands. Memoirs of the Carnegie Museum,
5: xii + 578 pages.
1912b. Some Results of an Ichthyological Reconnaissance of Colombia,
South America. Indiana University Studies, 10(8): 1-27.
1917. The American Characidae. Memoirs of the Museum of Comparative
Zoology, 43(1): 1-102.
1922. The Fishes of Western South America, Part 1. The Fresh-water
Fishes of Northwestern South America, Including Colombia,
Panama, and the Pacific Slopes of Ecuador and Peru, Together with
an Appendix upon the Fishes of the Rio Meta in Colombia. Memoirs
of the Carnegie Museum, 9:1-354.
67
68 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
Eigenmaim C.H., and W.R. Allen
1942. Fishes of Western South America: I, The Intercordilleran and
Amazonian Lowlands of Peru; II, The High Pampas, Bolivia, and
Northern Chile; with a Revision of the Peruvian Gymnotidae, and
of the Genus Orestias. 494 pages. Lexington: University of
Kentucky.
Eigenmann, C.H., and R.S. Eigenmann
1889a. Preliminary Descriptions of New Species and Genera of Characinidae.
The West Coast Scientist, 6:7-8.
1889b. A Revision of the Edentulous Genera of the Curimatinae. Annals
of the New York Academy of Sciences, 4:409-440.
Eigenmann, C.H., A. Henn, and C. Wilson
1914. New Fishes from Western Colombia, Ecuador, and Peru. Indiana
University Studies, 19:1-15.
Eigenmann, C.H., and C.H. Kennedy
1903. On a Collection of Fishes from Paraguay, with a Synopsis of the
American Genera of Cichlids. Proceedings of the Academy of
Natural Sciences ofPhiladephia, 55(2):497-537.
Eigenmann, C.H., and F. Ogle
1907. An Annotated l ist of Characin Fishes in the United States National
Museum and the Museum of Indiana University with Descriptions
of New Species. Proceedings of the United States National Museum,
33:1-36.
Eldredge, N., and J. Cracraft
1980. Phylogenetic Patterns and the Evolutionary Process. 349 pages.
New York: Columbia University Press.
Felsenstein, J., and E. Sober.
1986. Parsimony and Likelihood: An Exchange. Systematic Zoology
35(4):617-626.
Fernandez-Y6pez, A.
1948. Los Curimatidos (peces fluviales de Sur America): catalogo
descriptivo con nuevas adiciones genericas y especificas. Boletin
Taxonomico del Labor at or io de Pesqueria de Caiquire, 1:1 -86.
Fink, S.V., and W.L. Fink
1981. Interrelationships of the Ostariophysan Fishes (Teleostei). Zoologi-
cal Journal of the Linnean Society, 72(4):297-353.
Fowler, H.W.
1906. Further Knowledge of Some Heterognathus Fishes, Part I.
Proceedings of the Academy of Natural Sciences of Philadelphia,
58:293-351.
1911. New Fresh-water Fishes from Western Ecuador. Proceedings of the
Academy of Natural Sciences of Philadelphia, 63:493-520.
1913a. Fishes from the Madeira River, Brazil. Proceedings of the Academy
of Natural Sciences of Philadelphia, 65:517-579.
1913b. Curimatus spilurus Cope, a Wrongly Identified Characin. Proceed-
ings of the Academy of Natural Sciences of Philadelphia,
65:673-675.
1932. Zoological Results of the Matto Grosso Expedition to Brazil in 1931,
I: Fresh-water Fishes. Proceedings of the Academy of Natural
Sciences of Philadelphia, 84:343-377.
1940. A Collection of Fishes Obtained by Mr. William C. Morrow in the
Ucayali River Basin, Peru. Proceedings of the Academy of Natural
Sciences of Philadelphia, 91:219-289.
1941. A Collection of Freshwater Fishes Obtained in Eastern Brazil by
Dr. Rodolpho von Ihenng. Proceedings of Academy of Natural
Sciences of Philadelphia, 93:123-199.
1943. A Collection of Fresh-water Fishes from Colombia, Obtained
Chiefly by Brother Niceforo Maria. Proceedings of the Academy of
Natural Sciences of Philadelphia, 95:223-266.
1975. A Catalogue of World Fishes (XXIII). Quarterly Journal of the
Taiwan Museum, 28(3):277-4O2.
Gery, J.
1964a. Preliminary Descriptions of Seven New Species and Two New
Genera of Characoid Fishes from the Upper Rio Mela in Colombia.
Tropical Fish Hobbyist, 12(5):25-32, 41-48.
1964b. A Review of the Chilodinae, with a Key to the Species. Tropical
Fish Hobbyist, 12(9):5-10, 63-67, 9 figures.
1965. Notes on Characoid Fishes Collected in Surinam by Mr. H.P.
Pijpers, with Descriptions of New Forms. Bijdragen tot de
Dierkunde, 35:101-126.
1972. Corrected and Supplemented Descriptions of Certain Characoid
Fishes Described by Henry W. Fowler, with Revisions of Several
of their Genera. Studies on the Neotropical Fauna, 7:1-35.
1977a. Deux nouveaux cas de mimetisme chez les poissons: Curimatides
mimetiques d'Hemiodides (Characoides). Revue Frangais de
Aquariologie, 4:103-106.
1977b. Characoids of the World. 672 pages. Neptune City, New Jersey:
TFH Publications.
1984. The Fishes of Amazonia. In H. Sioli, editor, The Amazon: Limnology
and Landscape Ecology of a Mighty River and Its Basin, pages
333-370. Dordrecht: W. Junk Publishers.
Gill,T.
1858. Synopsis of the Fresh-water Fishes of the Western Portion of the
Island of Trinidad, W.I. Annals of the Lyceum of Natural History,
New York, 6:363-430.
1895. Notes on Characinoid Fishes with Ctenoid Scales, With a
Description of a New Psectrogaster. Proceedings of the United
Slates National Museum, 18:211-212.
Godoy, M.P. de
1975. Subordem Characoidei, Bacia do Rio Mogi Guassu. In Peixes do
Brazil, Volume 3: vi + pages 399-628. Piracicaba: Ed. Franciscana.
Goulding, M.
1981. Man and Fisheries on an Amazon Frontier. In H.J. Dumont, editor,
Developments in Hydrobiology, 4:137 pages, 37 figures. The Hague:
W. Junk Publishers.
Gray, J.E.
1854. Catalogue of Fish Collected and Described by L.T. Gronow, Now
in the British Museum. 1% pages. London.
Greenwood, PH., D.E. Rosen, S.H. Weitzman, and G.S. Myers
1966. Phyletic Studies of Teleostean Fishes with a Provisional Classifica-
tion of Living Forms. Bulletin of the American Museum of Natural
History, 131(4):339-456.
Gregory, W.K.
1933. Fish Skulls: A Study of the Evolution of Natural Mechanisms.
Transactions of the American Philosophical Society, new series,
23:76-481.
Gregory, W.K., and G.M. Conrad
1938. The Phylogeny of the Characin Fishes. Zoologica (New York),
23(4):319-360, figures 1-37.
Gronovius, L.T.
1763. Zoophylacii Gronoviani Fasciculus Primus Exhibens Animala
Quadrupeda, Amphibia, atque Pisces, etc. 136 pages. Lugduni
Balavorum.
Giinther, A.
1859. Second List of Cold-blooded Vertebrates Collected by Mr. Fraser
in the Andes of Western Ecuador. Proceedings of the Zoological
Society of London, 1859:402-420.
1864. Catalogue of the Fishes in the Collection of the British Museum.
Volume 5, xii + 455 pages. London.
1868. Descriptions of Freshwater Fishes from Surinam and Brazil.
Proceedings of the Zoological Society of London, for 1868:229-247.
1880. A Contribution to the Knowledge of the Fish-fauna of the Rio de
La Plata. Annals and Magazine of Natural History, series 5, 6:7-15.
Hecht, M.K., and J. Edwards
1977. The Methodology of Phylogenetic Inference at the Species Level.
In M.K. Hecht, PC. Goody, and B.M. Hecht, editors, Major
Patterns in Vertebrate Evolution, pages 3-51. New York: Plenum
Press.
NUMBER 471 69
Ilcnnig, W.
1950. Grundiziige einer Theorie der phylogenetischen Systematike. 370
pages. Berlin: Deutscher Zentralverlag.
1966. Phylogenetic Systematics. 263 pages. Urbana: University of Illinois
Press.
Hensel, R.H.
1869. The Freshwater Fishes of Southern Brazil. Archiv fur Naturges-
chichte (Wiegmann), 36(l):50-91.
Holmberg, E.L.
1891. Sobre algunos peces nuevos o poco conocidos de La Republica
Argentina. Revista Argentina de Historia Natural, 1:180-193.
Howes, G.J.
1976. The Cranial Musculature and Taxonomy of Characoid Fishes of the
Tribes Cynodontini and Characini. Bulletin of the British Museum
(Natural History), Zoology, 29:203-248.
International Commission of Zoological Nomenclature
1966. Opinion 772, Curimata Walbaum, 1792 (Pisces): Rejected as a
Generic Name and Placed on the Official List. Bulletin of Zoological
Nomenclature, 23(l):41-45.
Kapoor, B.G., and N.E. Evans
1975. The Gustatory System in Fish. Advances in Marine Biology,
13:53-108.
Kner, R.
1859. Zur Familie der Characinen, III: Folge der Ichthyologischen
Beitrage. Denkschriften der Akademie der Wissenschaften, Wien,
17:137-182.
Linnaeus, C.
1758. Systema Naturae. Edition 10, volume 1, 824 pages.
1766. Systema Naturae. Edition 12, volume 1, 532 pages.
Lowe-McConnell, R.H.
1975. Fish Communities in Tropical Freshwaters. 337 pages. New York:
Longman Publishing.
Lutken, C.F.
1874. Characinae Novae Brasiliae Centralis a Clarissimo J. Reinhardt in
Provincia Minas-Gerais circa Oppidulum Lagoa Santa in Lacu
Eiusdem Nominis, Flumine Rio das Velhas et Rivulis Affluentibus
Collectae, Secundum Caracteres Essentiales Breviter Descriptae.
Oversigt over del Kongelige Danske Videnskabernes Selskabs
Forhandlinger, 3:127-143.
Maddison, W.P., M.J. Donoghue, and D.R. Maddison
1984. Outgroup Analysis and Parsimony. Systematic Zoology, 33:83-103.
Mago-Leccia, F.
1970. Lista de los Peces de Venezuela. 241 pages. Caracas: Ministerio de
Agriculture y Cria.
Matthes, H.
1963. A Comparative Study of the Feeding Mechanisms of Some African
Cyprinidae (Pisces, Cypriniformes). Bijdragen lot de Dierkunde,
33:1-35.
McAllister, D.
1968. Evolution of the Branchiostegals and Classification of Teleostome
Fishes. National Museum of Canada, Bulletin 221 :l-239.
Meinken, H.
1933. Ueber einige in Letzter Zeit Eingefiihrte Fische II. Blatter fiir
Aquarien und Terrarienkunde, 44:71-73.
Miranda-Ribeiro, A. de
1937. Sobre uma colle^ao de vertebrados do nordeste Brasileiro. O Campo,
Rio de Janiero, 1:54-56.
Moriarty, D.J.W., J. Darlington, I.G. Dunn, CM. Moriarty, and M.P. Tevlin
1973. Feeding and Grazing in Lake Victoria. Proceedings of the Royal
Society of London, 184(B):299-319.
Miiller, J., and F. Troschel
1845. Horae Ichthyologicae, Beschreibung und Abbildung Neuer Fische;
Die Familie Characinen. 40 pages. Berlin.
Myers, G.S.
1927. Descriptions of New South American Fresh-water Fishes Collected
by Dr. Carl Temetz. Bulletin of the Museum of Comparative
Zoology, 68:107-135.
1929. On Curimatid Characin Fishes Having an Incomplete Lateral Line
with a Note on the Peculiar Sexual Dimorphism of Curimatopsis
macrolepis. Annals and Magazine of Natural History, series 10,
13:618-621.
Nelson, G.
1967. Epibranchial Organs in Lower Teleostean Fishes. Journal of
Zoology (London) 153:71-89.
1969. Gill Arches and the Phylogeny of Fishes with Notes on the
Classification of Vertebrates. Bulletin of the American Museum of
Natural History, 141(4):475-522.
1973. Relationships of Clupeomorphs, with Remarks on the Structure of
the Lower Jaw in Fishes. Zoological Journal of the Linnean Society,
53(supplement l):333-349.
Nomura, H., and C. Hayashi
1980. Caracteres meristicos e biologia do Saguiru, Curimatus gilberti
(Quoy and Gaimard, 1824), do Rio Morgado (Matao, Sao Paulo)
(Osteichthys, Curimatidae). Revista Brasileira de Biologia, 40(1):
165-176.
Nomura, H., and A.C. Taveira
1979. Biologia do Saguiru, Curimatus elegans Steindachner, 1874 do
Mogi Guacu, Sao Paulo (Osteichthys, Curimatidae). Revista
Brasileira de Biologia, 39(2):331-339.
Ortega, H., and R.P. Van
1986. Annotated Checklist of the Freshwater Fishes of Peru. Smithsonian
Contributions to Zoology, 437:25 pages.
Patterson, C.
1975. The Braincase of Pholodophorid and Leptolepid Fishes with a
Review of the Actinopterygian Braincase. Philosophical Transac-
tions of the Royal Society, series B, 269:275-579.
Paugy, D.
1984. Characidae. In J. Daget, J.-P. Gosse, and D.F.E. Thys van den
Audenaerde, editors, Check-List of the Freshwater Fishes of Africa,
pages 140-183. Paris: l'Office de la Recherche Scientifique et
Technique Outre-Mer.
Pearson, N.E.
1924. The Fishes of the Eastern Slopes of the Andes, I: The fishes of the
Riio Beni Basin Collected by the Mulford Expedition. Indiana
University Studies, 11 (64): 1 -83.
Pellegrin, J.
1908. Characinides Americains nouveaux de la collection du Museum
d'Histoire Naturelle. Bulletin du Museum National d'Histoire
Naturelle, Paris, 7:342-347.
1909. Characinides du Bresil rapportes par M. Jobert. Bulletin du Museum
National d'Histoire Naturelle, Paris, 15:147-153.
Puyo, J.
1943. Nouveaux poissons d'eau douce de la Guyane Franchise. Bulletin
de la Societe d'Histoire Naturelle, Toulouse, 78:141-149.
Quoy, J.R., and P. Gaimard
1824. Zoologie, Poissons. In Voyage autour du monde, execute sur les
corvettes de SM "L'Uranie"e "LaPhysicienne" Pendant les Annes
1817-1820. Pages 192-401. Paris.
Regan, C.T
1911. The Classification of the Teleostean Fishes of the Order Ostario-
physi, I: Cyprinoidea. Annals and Magazine of Natural History,
series 8, 8:13-32.
Reid, G. McG.
1982. The Form, Function and Phylogenelic Significance of the Vomero-
Palatine Organ in Cyprinid Fishes. Journal of Natural History,
16:497-510.
70 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
Ringuelet, R.A., R.H. Aramburu, and A. Alonso de Aramburu
1967. Los peces Argentinos de Agua Dulce. 602 pages. LaPlata: Comision
de Investigacion Cientifica.
Risso, FJ.J., and E.L. Sanchez
1964. Una nueva subespecie de Curimatinae. Notas del Museo de Ciencias
Naturales del Chaco, Resistencia, 1:6-10.
Roberts, T.R.
1966. Description and Osteology of Lepidarchus adonis, a Remarkable
New Characoid Fish from West Africa. Stanford Ichthyological
Bulletin, 8:209-227.
1969. Osteology and Relationships of Characoid Fishes, Particularly the
Genera Hepsetus, Salminus, Hoplias, Ctenolucius, and Acestrorhyn-
chus. Proceedings of the California Academy of Sciences, 36(15):
391-500, figures 1-68.
1973. Osteology and Relationships of the Prochilodontidae, a South
American Family of Characoid Fishes. Bulletin of the Museum of
Comparative Zoology, 145(4):213-235, figures 1-29.
1974. Osteology and Classification of the Neotropical Characoid Fishes
of the Families Hemiodontidae (including Anodontinae) and
Parodontidae. Bulletin of the Museum of Comparative Zoology,
146(9):411-472, figures 1-78.
Rodriquez, G.
1973. El Sistema de Maracaibo, Biologia y Ambiente. 395 pages. Caracas:
Instituto Venezolano de Investigaciones Cientificas.
Saint-Paul, U., and P. Bayley
1979. A situaijao da pesca na Amazonia Central. Acta Amazonica,
9(4):109-114.
Santos, G.M. dos, M. Jegu, and B. de Merona
1985. Catdlago de peixes commercisis do baixo rio Tocantins. 85 pages.
Manaus: Eletronorte.
Smith, N.
1981. Man, Fishes and the Amazon. 180 pages. New York: Columbia
University Press.
Spix, J.B. von, and L. Agassiz
1829. Selecta Genera et Species Piscium Quos in Itinere per Braziliam
Annis 1817-20 Perecato Collega J?. de Spix. 138 pages. Munich.
Steindachner, F.
1874. Die Siisswasserfishe des Siidostlichen Brasiliens. Sitzungsberichte
der Akademie der Wissenschaften. Wien, 70:499-538.
1876. Ichthyologische Beitrage (V.), II: Uber einige neue Fisharten
insbesondere Characinen und Siluroiden aus dem Amazonenstrome.
Sitzunsberichte der Akademie der Wissenschaften, Wien, 74:49-
240.
1878. Fischfauna des Magdalenen Stromes. Denkschriften der Akademie
der Wissenschaften, Wien, 39:47-69.
1879. Beitrage zur Kenntniss der Flussfische Sudamerika's. Denkschriften
der Akademie der Wissenschaften, Wien, 41:161-172.
1882a. Beitrage zur Kenntniss der Flussfische Sudamerika's. Denkschriften
der Akademie der Wissenschaften, Wien, 46:1-44.
1882b. Ichthyologische Beitrage (XII.). Sitzungsberichle der Akademie der
Wissenschaften. Wien, 84:61-82.
1908. Ueber zwei neue Siluroiden und zwei Curimatus-Arten, sowie uber
eine Varietal con Ancistrus vittatus aus dem Amazonasgebeite
Innerhalb Brasiliens. Anzeiger der Akademie der Wissenschaften,
Wien. 45:163-168.
1910. Noliz uber einige neue Characinenarten aus dem Orinoco und dem
Oberen Surinam. Anzeiger der Akademie der Wissenschaften, Wien,
47:265-270.
1911. Uber vier neue Siluroiden und Characinen aud dem Amazonasgebe-
ite und von Ceara aus der Sammlung der Museums Goldi in Para.
Anzeiger der Akademie der Wissenschaften, Wien, 48:324-331.
1914. Ueber eine neue Basilianische Curimalus-An:Curimatus semior-
natus n. sp. Anzeiger der Akademie der Wissenschaften, Wien,
48(12):262-263.
1917. Beitrage zur Kenntniss der Flussfische Sudamerika's. Denkschriften
der Akademie der Wissenschaften, Wien, 93:15-106.
Swofford, D.L
1985. PAUP: Phylogenetic Analysis Using Parsimony, Version 2.4.
Champaign: Illinois Natural History Survey. [Computer program.]
Taylor, W.R., and G.C. Van Dyke
1985. Revised Procedures for Staining and Clearing Small Fishes and
Other Vertebrates for Bone and Cartilage Study. Cybium, 9(2):107-
119.
Van, R.P.
1979. Anatomy, Relationships and Classification of the Families Cithar-
inidae and Distichodontidae (Pisces, Characoidea). Bulletin of the
British Museum (Natural History), Zoology, 36(5) :261-344.
1982a. Systematics of the Neotropical Characoid Genus Curimatopsis
(Pisces: Characoidei). Smithsonian Contributions to Zoology,
373:28 pages, 21 figures.
1982b. Curimatopsis myersi, a New Curimatid Characiform Fish (Pisces:
Characiformes) from Paraguay. Proceedings of the Biological
Society of Washington, 95(4):788-792.
1983. Phylogenetic Relationships of the Families Curimatidae, Prochi-
lodontidae, Anostomidae, and Chilodontidae (Pisces: Characi-
formes). Smithsonian Contributions to Zoology, 378:60 pages, 41
figures.
1984a. Systematics of the Neotropical Characiform Genus Potamorhina
(Pisces: Characiformes). Smithsonian Contributions to Zoology,
400:36 pages, 17 figures.
1984b. Two New Fish Species of the Genus Curimata (Pisces: Curimatidae)
from Venezuela. Acta Biologica Venezuelica, 11(4):27-43.
1985. A New Species of Bivibranchia (Pisces: Characiformes) from
Surinam, with Comments on the Genus. Proceedings of the
Biological Society of Washington, 98(2):507-518.
1987. Two New Species of Curimatid Fishes (Ostariophysi: Characi-
formes) from Rio Grande do Sul, Brazil. Proceedings of the
Biological Society of Washington, 100(3):603-609.
In press. Systematics of the Neotropical Characiform Genus Curimata
Bosc (Pisces: Characiformes). Smithsonian Contributions to Zool-
ogy, 474.
Van, R.P., and R.M.C. Castro.
1988. Prochilodus stigmaturus Fowler, Reassigned to the Curimatidae
from the Prochilodontidae, with Comments on Other Nominal
Curimatid and Prochilodontid Species Treated by Fowler. Copeia
1988(3):777-780.
Vari.R.P.andJ. Gery
1985. A New Curimatid Fish (Characiformes: Curimatidae) from the
Amazon Basin. Proceedings of the Biological Society of Washing-
ton. 98(4):1030-1034.
Vari, R.P., and H. Nijssen
1986. Curimata punctata, a New Uniquely Pigmented Species of
Curimatid from the Marowijne River Basin of Surinam and French
Guiana (Pisces, Characiformes). Beauforlia, 36(4):51-55.
Wasserzug, R.
1972. The Mechanisms of Ultraplanktonic Entrapment in Anuran Larvae.
Journal of Morphology, 137(3):279-288.
Watrous, L.E., and W.D. Wheeler
1981. The Out-Group Comparison Method of Character Analysis.
Systematic Zoology, 30:1 -11.
Weitzman, S.H.
1962. The Osteology of Brycon meeki, a Generalized Characid Fish, with
an Osteological Definition of the Family. Stanford Ichthyological
Bulletin, 8(l):l-77, figures 1-21.
1964. Osteology and Relationships of South American Characid Fishes
of Subfamilies Lebiasinidae and Erythrinidae with Special Refer-
ence to Subtribe Nannostomina. Proceedings of United States
National Museum, 3499:127-170, figures 1-10.
NUMBER 471 71
Weitzman, S.H., and S.V. Fink
1985. Xenurobryconin Phylogeny and Putative Pheromone Pumps in
Glandulocaudine Fishes (Teleostei: Characidae). Smithsonian Con-
tributions toZoology, 421:121 pages, 81 figures.
Weitzman, S.H., and W.L. Fink
1983. Relationships of the Neon Tetras, A Group of South American
Freshwater Fishes (Teleostei, Characidae), with Comments on the
Phylogeny of New World Characiforms. Bulletin of the Museum of
Comparative Zoology 150(6):339-395, figures 1-30.
Weitzman, S.H., and R.P. Vari
1988. Miniaturization in South American Freshwater Fishes: An Overview
and Discussion. Proceedings of the Biological Society of Washing-
ton, 101(2):444-465.-
Whitley, G.P.
1953. Studies in Ichthyology. Records of the Australian Museum,
23:133-158.
1954. New Locality Records for Some Australian Fishes. Proceedings of
the Royal Society of New South Wales, 1952-1953:23-30.
Wiley, E.O.
1981. Phylogenetics. The Theory and Practice of Phylogenetic Systema-
tics. vi + 439 pages. New York: John Wiley and Sons.
Winterbottom, R.
1974. A Descriptive Synonymy of the Striated Muscles of the Teleostei.
Proceedings of the Academy of Natural Sciences of Philadephia,
125:225-317.
1980. Systematics, Osteology and Phylogenetic Relationships of Fishes
of the Ostariophysan Subfamily Anostominae (Characoidei, Anosto-
midae). Life Sciences Contribution, Royal Ontario Museum,
123:1-112, figures 1-68.
A U.S. GOVERNMENT PRINTING OFFICE: 1988-181-717/60028

REQUIREMENTS FOR SMITHSONIAN SERIES PUBLICATION
Manuscripts intended for series publication receive substantive
review (conducted by their originating Smithsonian museums or
offices) and are submitted to the Smithsonian Institution Press
with Form SI-36, which must show the approval of the appropriate
authority designated by the sponsoring organizational unit. Re-
quests for special treatment?use of color, foldouts, case-bound
covers, etc.?require, on the same form, the added approval of
the sponsoring authority.
Review of manuscripts and art by the Press for requirements
of series format and style, completeness and clarity of copy, and
arrangement of all material, as outlined below, will govern, within
the judgment of the Press, acceptance or rejection of manuscripts
and art.
Copy must be prepared on typewriter or word processor,
double-spaced, on one side of standard white bond paper (not
erasable), with 1 VA" margins, submitted as ribbon copy (not
carbon or xerox), in loose sheets (not stapled or bound), and
accompanied by original art. Minimum acceptable length is 30
pages.
Front matter (preceding the text) should include: title page
with only title and author and no other information, abstract page
with author, title, series, etc., following the established format;
table of contents with indents reflecting the hierarchy of heads in
the paper; also, foreword and/or preface, if appropriate.
First page of text should carry the title and author at the top
of the page; second page should have only the author's name
and professional mailing address, to be used as an unnumbered
footnote on the first page of printed text.
Center heads of whatever level should be typed with initial
caps of major words, with extra space above and below the head,
but no other preparation (such as all caps or underline, except for
the underline necessary for generic and specific epithets). Run-in
paragraph heads should use period/dashes or colons as neces-
sary.
Tabulations within text (lists of data, often in parallel columns)
can be typed on the text page where they occur, but they should
not contain rules or numbered table captions.
Formal tables (numbered, with captions, boxheads, stubs,
rules) should be submitted as carefully typed, double-spaced copy
separate from the text; they will be typeset unless otherwise
requested. If camera-copy use is anticipated, do not draw rules
on manuscript copy.
Taxonomic keys in natural history papers should use the
aligned-couplet form for zoology and may use the multi-level
indent form for botany. If cross referencing is required between
key and text, do not include page references within the key, but
number the keyed-out taxa, using the same numbers with their
corresponding heads in the text.
Synonymy in zoology must use the short form (taxon, author,
year: page), with full reference at the end of the paper under
"Literature Cited." For botany, the long form (taxon, author,
abbreviated journal or book title, volume, page, year, with no
reference in "Literature Cited") is optional.
Text-reference system (author, year:page used within the text,
with full citation in "Literature Cited" at the end of the text) must
be used in place of bibliographic footnotes in all Contributions
Series and is strongly recommended in the Studies Series
"(Jones. 1910:122)" or "...Jones (1910:122)." If bibliographic
footnotes are required, use the short form (author, brief title, page)
with the full citation in the bibliography.
Footnotes, when few in number, whether annotative or biblio-
graphic, should be typed on separate sheets and inserted imme-
diately after the text pages on which the references occur. Exten-
sive notes must be gathered together and placed at the end of
the text in a notes section.
Bibliography, depending upon use, is termed "Literature Cited,"
"References," or "Bibliography." Spell out titles of books, articles,
journals, and monographic series. For book and article titles use
sentence-style capitalization according to the rules of the lan-
guage employed (exception: capitalize all major words in English).
For journal and series titles, capitalize the initial word and all
subsequent words except articles, conjunctions, and prepositions.
Transliterate languages that use a non-Roman alphabet according
to the Library of Congress system. Underline (for italics) titles of
journals and series and titles of books that are not part of a series.
Use the parentheses/colon system for volume (number):
pagination: "10(2):5-9." For alignment and arrangement of ele-
ments, follow the format of recent publications in the series for
which the manuscript is intended. Guidelines for preparing bibli-
ography may be secured from Series Section, SI Press.
Legends for illustrations must be submitted at the end of the
manuscript, with as many legends typed, double-spaced, to a
page as convenient.
Illustrations must be submitted as original art (not copies)
accompanying, but separate from, the manuscript. Guidelines for
preparing art may be secured from Series Section, SI Press. All
types of illustrations (photographs, line drawings, maps, etc.) may
be intermixed throughout the printed text. They should be termed
Figures and should be numbered consecutively as they will
appear in the monograph. If several illustrations are treated as
components of a single composite figure, they should be desig-
nated by lowercase italic letters on the illustration; also, in the
legend and in text references the italic letters (underlined in copy)
should be used: "Figure 9b." Illustrations that are intended to
follow the printed text may be termed Plates, and any components
should be similarly lettered and referenced: "Plate 9b." Keys to
any symbols within an illustration should appear on the art rather
than in the legend.
Some points of style: Do not use periods after such abbrevi-
ations as "mm, ft. USNM, NNE." Spell out numbers "one' through
nine" in expository text, but use digits in all other cases if possible.
Use of the metric system of measurement is preferable; where
use of the English system is unavoidable, supply metric equiva-
lents in parentheses. Use the decimal system for precise meas-
urements and relationships, common fractions for approximations.
Use day/month/year sequence for dates: "9 April 1976." For
months in tabular listings or data sections, use three-letter abbre-
viations with no periods: "Jan, Mar, Jun," etc. Omit space between
initials of a personal name: "J.B Jones."
Arrange and paginate sequentially every sheet of manu-
script in the following order: (1) title page. (2) abstract, (3) con-
tents, (4) foreword and/or preface, (5) text, (6) appendixes, (7)
notes section, (8) glossary, (9) bibliography (10) legends, (11)
tables. Index copy may be submitted at page proof stage, but
plans for an index should be indicated when manuscript is sub-
mitted.
??
,111/.,