Cover: Hyphessobrycon serpae. Photo by H.J. Richter.

Quotations of reasonable length in scientific, non-com-

mercial pUblications are allowable without charge and
without permission from the publisher.
ISBN 0-87666-458-3
~~iy
Printed in the U.S.A.
@Copyrlght 1977 by T.F.H. Publlcatlons, Inc. Ltd.
Dislribuled in lhe U.S.A. by T.F.H. Publicolions, Inc., 211 Wesl Sylvonio Avenue, P.O. Box 27,
Neplune Cily, N.J. 07753; in Englond by T.F.H. (GI. Brilain) Lld., 13 Nulley Lone, Reigole, Surrey;
in Conodo 10 lhe book slore and library Irdde by Clarke, Irwin &Compony, Clarwin House, 791 SI.
Cloir Avenue Wesl, Toronlo 10, Onlario; in Canodo 10 lhe pel Irade by RolI C. Hagen Lld., 3225
Sortelon SIreel, Monlreol382, Quebec; in Soulheosl Asia by Y.W. Ong, 9 Lorong 36 Geylang,
Singapore 14; in Auslralia and lhe soulh Pacilic by Pel Imporls Ply. Lld., P.O. Box 149, Brookvole
2100, N.S.W., Auslralio. Published by T.F.H. Publicalions, Inc. Lld., The Brilish Crown Colony 01
Hong Kong.
2
-----------..
PREFACE
This book, devoted to the most diversified group' of American and
African fresh-water fishes, has two goals. First, to provide the general
reader, who is an aquarist more often than not, with as many good illustra-
tions of aquarium type characoid fishes as possible. This kind of pictorial
guide will enable him, it is hoped, to identify at first sight a good portion of
the aquarium species. At the same time, the non-specialized reader will get
an idea of the position of the species in the classification(genus, family and
so on) and, if he wants, can refer to the text for more systematic and,
occasionally, biological data, using the index of Latin names at the end of
the volume.
The second aim of the book is to give to zoologists as well as to advanced
aquarists a scientific account of our present knowledge of the group, which
has not been reviewed as a whole for balf a century. This account has been
made as complete and reliable as possible, with a minimum of words and a
bibliography reduced to a few recent works, but it is by no means to be
taken as a true revision of the suborder.
In order to classify and to describe all, or aImost all, known species and
enable curators of museum cllections and other interested specialists to
identify them with reasonable accuracy, dichotomous tables, also called
identification keys or dichotomous keys, are presented along with a
discussion of the various species and their particular characteristics.
The author is aware that such a system causes a kind of psychological
block to certain persons. These readers are asked to content themselves with
the illustrations and general statements regarding the groups, which should
be quite sufficient in most cases. Others, it is hoped, will make the simple
effort to enter the field of "yes or no" reasoning, which is basically the
computer's, as well as our own brain's, way of thinking. To master the use
of a key, which only takes a few minutes, the principIe is as follows: one is
asked, step by step (a, b, c, etc.) a question to which the answer is either yes
or no. If the organism to be identified has such or such a feature (answer
yes), one goes to the next step. If not (answer no), one goes to the corres-
ponding branch of the altemative (represented by aa, bb, cc, etc.), which is
more or less remote in the key, until the identification is achieved.
II
L!:cb ;. >:c., .':':-;1 Py -Danlel
c. Po.<;;t..d: 'i"". :-~'P~!1
69,000 - Marwu - Amo.o - B,ulJ
3
Actually, this will probably not always be the case, and the reader must
be prepared for some falseleads. The impossibility of identifying a particular
species may be due to the foJlowing: certain keys are surely imperfect,
suffering from some defect wbich can only be detected after several tests;
certain exceptions t9 the presence or absence of a character may cause
difficulties, although they should be mentioned in the key when known;
and the whole group is still imperfectly known and it may be, cbiefty if the
species is small and comes from some remote area, that the unidentifiable
specimen has never been scientificallydescribed (a so-called new species).
The characoids now number more tPan 1200 species (inc1uding200 or so
Mrican ones), and, based on past experience, in every large collecting
sample 5-10% of the species can be expected to be new.
ACKNOWLEDGMENTS
The author wishes to extend bis thanks to the persons who have gready
contributed to the book:
Dr. Warren Burgess, who had the difficult task of making the author's
language readable by English-speaking people, and the people whose
remarkable photographs provide the indispensable illustrations, principally
Dr. H. R. Axelrod and the late Harald Schultz, in whose memory this book
is dedicated.
4 5
-~
"""I
Table of Contents
('
Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Acknowledgments 5
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Part 1: The African FamUies. . . . . .. . . . . .. . . . . . . . . . . . . . . . . .. .17
Key to the African Families . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Family Hepsetidae (African pike-like characoids) . . . . . . . . . . .17
Family Alestidae (African tetras). . . . . . . . . . . . . . . . . . . . . . . . .18
Subfamily Alestinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
Subfamily Hydrocyninae . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54
Family Citharinidae (citharinids) . . . . . . . . . . . . . . . . . . . . . . . .54
Subfamily Citharininae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55
Subfamily Distichodinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58
Subfamily Ichthyborinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . .82
Part 2: The American Characoids other than the Characidae. . . . .95
Key to the American Families . . . . . . . . . . . . . . . . . . . . . . . . . . . . .95
Family Erythrinidae (trahiras) . . . . . . . . . . . . . . . . . . . . . . . . . .98
Family Ctenoluciidae (South American pike-characoids). . . .103
Family Crenuchidae. . . . . . . ; _. . . . . . . . . . . . . . . . . . . . . . . . . 106
Family Characidiidae (South American "darters") . . . . . . . . .111
Subfamily Characidiinae. . . . . . . . . . . . . . . . . . . . . . . . . . . .114
SubfamilyElachocharacinae .119
Subfamily Geisleriinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Family Lebiasinidae (pencilfishesand pyrrhulinins) . . . . . . . .123
Subfamily Lebiasininae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Subfamily Pyrrhulininae . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Family Anostomidae (headstanders and leporins) . . . . . . . . . .150
Subfamily Leporellinae . . . . . . . . . . . . . . ;' . . . . . . . . . . . . .151
Subfamily Anostominae. . . . . . . . . . . . . . . . . . . . . . . . . . . . .151
Family Hemiodidae (hemiodids and their relatives) . . . . . . . .190
Subfamily Hemiodinae .190
Subfamily Parodontinae 202
Family Curimatidae (curimatas). .210
Subfamily Chilodinae 211
Subfamily Prochilodinae . . . . . . . . . . . . . . . . . . . . . . . . . . . .214
Subfamily Curimatinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
7
Subfamily Anodinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
Family Gasteropelecidae (hatchetfishes) 243
Family Serrasalmidae (pacus, silver dollars and piranhas) . . .251
Subfamily Myleinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .251
Subfamily Serrasalminae . . . . . . . . . . . . . . . . . . . . . . . . . . . .275
Subfamily Catoprioninae. . . . . . . . . . . . . . . . . . . . . . . . . . :. 294
Part 3: The American Tetras and Characins (Family Characidae)
.295
Subfamily Agoniatinae. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
Subfamily Rhaphiodontinae . . . . . . . . . . . . . . . . . . . . . . . . . 299
Subfamily Characinae . . . . . . . . . . . . . . . . . . . . . . . . . .. . . .302
Subfamily Bryconinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
Subfamily Clupeacharacinae 346
Subfamilies Paragoniatinae and Aphyocharacinae . . . . . . .347
SubfamilyGlandulocaudinae .351
Subfamily Stethaprioninae . . . . . . . . . . . . . . . . . . . . . . . . . . 366
Subfamily Tetragonopterinae . . . . . . . . . . . . . . . . . . . . . . . .371
Subfamily Rhoadsiinae 538
Subfamily Cheirodontinae and allied genera . . . . . . . . . . . . 543
Undescribed or unidentified species . . . . . . . . . . . . . . . . . . . . . . . . .609
Glossary of technical terms . . . . . . . . . . . . . . . . . . . . . . . . . .611
Bibliography I: Africa 621
Bibliography 11: South America . . . . . . . . . . . . . . . . . . . . . . . . . . . .642
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 647
INTRODUCTION
Onee upon a time, an Mrican fresh-water fish decided to beeome a
eharaeoid.
This happened long ago, in the Mesozoie (70-175 millions of years?)
when eonditions were favorable. Volcanism was intense, working profound
modifications of the earth. Our fish lived iJ) a mythical land, called Gond-
wanaIand, formed by the united continents of Mrica and South America.
They were soon to be separated, owing to a gigantie break which had just
started, forming a ehain of great lakes. A similar process is actualIyin action
in 'Mrica today; the new break starts with the Red Sea and is prolonged by
the Rift VaIley forming the Great Lakes of Africa (Tanganyika, Malawi,
Victoria, ete.).
Sueh lakes are the most suitable plaees to alIowwhat is calIed explosive
evolution to oecur because they offer the greatest ecologica1opportunities
to the species. These species are able to oecupy a multitude of empty spots
sueh as rocks, sand, grass-beds, bottom, surfaee, and so on, or potentiaI
so-calIed eeological niehes. Another possible cause of rapid and active
speeiation in Great Lakes, as eontrasted with the "normal" evolution in
adjaeent rivers, may be the laek of important fish predators (sueh as
Hydrocynusor Lates in certain Mrican lakes) whieh, when present, restrain
the diversification of the species.
The ambitious fish mentioned at the beginning of the story (which is not
a fairy-tale but a hypothesis based on eertain welI established facts), was
probably a species resembling what we now calI a gonorynchiform. That
smalI group of partIy fresh-water fishes, mostly Mrican, with Kneria,
Cromeria and others, seems indeed to have some "pre-ostariophysine"
features.
An ostariophysine fish (eharaeoids were apparently the first ostario-
physans, although certain authors believe that the catfishesare more ancient)
is eharacterized by at least two features: a chain of ossic1es,calIed Weber's
ossic1es,uniting the gas-bladder with the internal ear and permitting the
hearing of high-frequency sounds; and a chemical alarm system, telling the
sehooling fish that one of its eongeners has been attacked by a predator.
These two adaptations are supposed to be useful to fishesliving in schools in
fresh-water biotopes (which is the case in inost characoids and cyprinoids).
8
9
.. e <t ~ 61
~ ::tLl) -~ G>u
- =' ~_Q)o~ _~ E
" - Q.u ,,- :=.~ ~ Q) ~ C'f")
"0._ o.._~u..a_ Q) o. ..c.
Q. ~ ~ u 0Z;..! ...Q s... ~ ~ Q. o -+- C
Q.cnOuUL-Q)dU') o...c ~ -+- 11)
0"""000.1000 <:1 cc~
L-Q) ,,""co__Q. ""'111 U ,-_O
~~~o...tl CI Q. nU L-
61 I ~
."
o
C
61
L
O-
.,
o
..
61
....
O-
ti
....
~
..
o
ti
..
-
C
~
..
o
ti
..
O-
"
~ .,
..
o
."
o
e
L
....
61_-
.2 ::I
.- n
-f ::I
2 i: =
C ._
ti )(
ti
e
61
..
o-
- >.
- .. >.
ti ti ..
_ ti
,:=.....
.. )( c
o ti 61
ti e'"
..
-
C
." ." ."
' ' ' C\J ~
O'iCO'i-"'3
~~ ~~~
Q) Q. c ...c '-
V) ""' "
o- ti o- o
o L.. o o
o U) ~
u o. c c
61 e.-
L
.. ti
..
61
....
ti
10
Recent investigations have shown that most, if not ali, of the Kneria-like
recent fishesdo possess rudiments of these structures. They have specialized
cervical vertebrae (the first three) as if they were in the process of differen-
tiating ossicles (which is achieved in the characoids), and they also have a
typical fright reaction if a wounded fish of their kind (or a cyprinid) is
introduced into a school.
With ali the above cited conditions brought together, our Kneria-like
hypothetical procharacoid was roughly in the same position as a cichlid was,
rather recently, in a great Mrican lake, that occupied alI potential (empty
or not completely exploited) niches and gave rise to the famous, extremely
diversified, recent fauna ofthe Great Lakes. It is likelythat the procharacoid
literally underwent explosive radiation with such a spread of adaptations
that there are no other such examples in the vertebrates except for the
Australian marsupiaIs.
This evolution was probably possible thanks to the general organization,
with several generalized features, which characterizes a characoid. In this
group, therefs usually an adiposefin and, almost always, weli developedscales.
Teeth are alwayspresent, at leastin thepost-Iarva. Thereis no rostralbone,no
preethmoid, and the prevomer is not toothed and is direct/y articulated with
the palatine head and lateral ethmoid. Final/y, the Weberian ossiclesare
complete,with the tripusfixed on the body 01the third vertebraby means 01a
vibrating lamelia, in contrast to ali other ostariophysans.
Once the continents were separated, the above defined characoids evolved,
for their own part, on each side of the newly-formed Atlantic OceanoIt was
very probably impossible for these strictly so-called primary fresh-water
fishes to survive in marine water, except for short periods of time or short
distances.
It is believed (but not by ali zoologists) that in Mrica they soon gave rise
to the cyprinids, of which the primitive stock could have been Barilius-like
fishes (?), which in turn radiated and became ardent competitors of their
own parents (the descendants of the procharacoids) in the very same
biotopes. These barbs emigrated to Asia and North America by the
northem route, while the characoids apparently never did (Astyanax
lasciatus reached Texas by a different route, the southem one through
Central America).
In America, the characoids gaverise to the highly specializedgymnotoids
after the separation of the continents. Gymnotoids never developed in Mrica,
where their niches are occupied by knife-fishes and mormyroids.
Neither the gymnotoids nor the catfishes (of unknown origin), specialized
very early, were such competitors as were the barbs in Mrica. The
American (Neotropical) characoids were thus free to diversify in almost ali
possible directions. In addition, the drift of the continents had caused the
ti
.,
..
o
."
11
--.:.-
N
co
c
T'"
I:
as
E
N
-
Q)
...
Q)
-
o
as
.&;
(J
'C
Q)
Q)
I:
Q)
CI
as
-
ti
o
>.
L
:;
o
o-
>.
ti)
L
Q)
.&;
-
-
o
Q)
">
Q)

..J
""
upheaval of the Andes, as if the westem coast of South America had abutted
against a static block in its course to the west. This in turn caused the
formation of the Amazon basin, the greatest of the world, which offered the
best opportunities to a young, opportunistic group of fishes.
These two reasons explain why the Neotropica1 characoid fauna repre-
sents, in number of species, about 80% of the total, and the African
("Ethiopian") fauna consequently only 20%, with 11 and 3 families res-
pectively, according to certain modem authors.
On each side of the Atlantic Ocean, similar ecological pressures deter-
mined a parallel evolution. Certain evolutionary convelgences are worth
mentioning:
The African pike-characoidHepsetushas its counterpart in the
American Acestrorhynchus.
The large African alestids (except Hydrocynus) strongly parallel the
American Bryconinae, whereas the small alestids are scarcely discemable
from the numerous American tetras.
Hydrocynus, a modification of an alestid toward predation, has dose
similarities with the Neotropical Hoplias or Erythrinus.
Most of the Citharinidae are convergent with certain Hemiodidae and
Curimatidae (two rather dosely related American families), whereas
some others, very small and bottom-living fishes(Nannocharaxand allies),
are strikingly similar to the Neotropical Characidiidae.
1n the above cited examples, only knowledge of the origin, or the study
of anatomical details, can tell the non-specialist whether he is dealing with
an African or an American line, which diverged from each other so many
millions of years ago. Hepsetus, the most primitive characoid according
to T. Roberts (forming the family Hepsetidae), has an accessoryectoptery-
goid bone and a canal into the supraoperde. These characters are not shared
by any Neotropical characoid, even the Ctenoluciidae which are apparently
the dosest group. The citharinids have ctenoid scales(except for one genus),
numerous ventral rays, pharyngeal "bags" and other features which do not
exist (at least together) in American families. Finally, the alestids differ
from their dosest American relatives, the Characidae, in having no rhino-
sphenoid, a bony tube endosing the olfactory nerve, a toothless, paddle-like
maxilla, and other characters.
1t has been pointed out above that the evolution of the characoids, chiefty
of the Neotropical ones, resembles that of the more recent cichlids from
certain African lakes. Both groups, each with a different time scale however,
entered new biotopes and exploited all of their possibilities.
1ndeed, mos! of the adaptations enumerated by Fryer and Iles (1972)
conceming the African cichlids are found in the characoids. They practice
commensalism, mimicry,shoaling (alsocalled schooling),migration, burrow-
12
13
CII
....
e:
o
a.
c
11)
. I-
11)-0
>.CII
"J::. ...-
a. I-
-O
o CII
l-
a.>
-c
C-'=
-
11)
1-.....
.: CII
c .
11)
-'= 11)
a. 11)
>.
..... I-
-'=
.....
CII
a.
N
>
o c
o 11)
a. I-
CO
e:
C-O
O)
CII C
I-
CII
....
U
I-
c..... e
.- .....
a. I-
ai
-c
CII
CII
.., I-
> E
a.-'=
-'=
N
.....
-
o+-
\+-
....
o
C .-
11)....
"-
CII
11)
CD
o.....
-
c:
.... o
-
ii
11)
"

o
f!!
ai
.J::.
o
"
Q)
e
I-
-
::I
::I i
- I-
"-
C
.....
CD
11)
o I-
e:
c:
::I
11) ::I
CII
CD
a.
e:
CII
u
ti)
CII -
ai
a. a.
11)
-
::I
o
'"
fi)
:J
)(
li)
-
CII
o
f!!
ii
I
I
I
ai
Q.
e
....
Q.
C
e
e
o
e
ai
u
::I
e
e
::I
c:
I-
::I
::I
11)
I-
ai
11)
::I
o+-
.....
.- I- CII
e:
".:;
11)
-'= c
I-
.8
c
::I
a.
-
u
.....
CII
I-
oN
u
c
c
c
e:
CD
::I
-
U
I-
CII
CII
U
U
I-
a.e
u
e:
11) (\I
.....
-
::I
e:
c';:
I- e:
CIICII
0+- U
C
- --
-1
Alestidae, Serrasalmidae, Crenuchidae and even in one curimatid genus,
whereas care of the young appeared in at least three lines, certain Erythrini-
dae and Serrasalmidae, and certain Pyrrhulininae in the family Lebiasinidae.
The above enumeration of their intricate adaptations helps to explain
why the systematics of the characoids has long been, and stilI is, a difficult
task. They were recognized as a homogeneous group only in 1844 by two
German ichthyologists, MlIer and Troschel. The Viennese Kner and his
student Steindachner (between 1858and 1915) folIowedby accurately des-
cribing a number of species without paying much attention to their c1assi-
fication. It remained for one of the best ichthyologists of alI times, Carl H.
Eigenmann, to undertake, in the U.S., a natural c1assificationof the South
American groups. His principal monographs were issued between 1910and
1927. He was folIowedby his student, George S. Myers, the still very active
father of the "characologists," who in turn handed the torch down to his
own students, J. Bhlke, S. Weitzman and several others, forming a chain
of interest which is presently stilI growing.
Recentlya series of anatomical studies by S. Weitzman, T. Roberts, and
Others helped specify the systematic position of several critica1 groups,
which led to the recognition of many families within the suborder (a division
which could never have been accepted in Eigenmann's time) and revealed
the importance, as welI as the complexity, of the characoids, presently
considered as a key-group among the teleosts.
The first authoritative division ofthecharacoids into several families (16)
was done by S. Weitzman, in Greenwood et aI. (1966). The arrangement of
the families in the present book takes into account that c1assification,as
modified by Gry (1972), with a few alterations from the 1972 order.
As compared with Weitzman's c1assification, the present one is as
folIows:
ing in sand, and so on. They have developed three means of communication,
namely sounds, chemicals, and recognition marks (mostly on humeral
region .and caudal pedunc1e, as welI as on dorsal and adipose fins). They
have, in certain cases, elaborated or evolved breeding habits such as the
care of the young as practiced by the piranhas. Finally, they have developed
an array offeeding habits, which is largely comparable to that ofthe cichlids,
inc1uding scale eating or fin biting, as well as other adaptations which are
almost unique among teleosts, such as flight (hatchet-fishes) or group-
predation (piranhas).
Owing to the millions of years of evolution at the disposal 0f the
characoids, these adaptations ocurred more than once and in different
phylogenetic lines. This resulted in an extremely confusing complexity of
groups. It is often difficult to telI from a character shared by different forms
the role of true phylogeny (common origin) and of convergency (different
origin). That is how a generalized body-shape, like that of a salmon or a
trout for example, is to be seen at the same time in Brycon, Salminus and
Astyanax, which originated from three different evolutionary lines
(although in the same family).
Evolution of a predatory existence, that is predation mostly on fishes,
appeared at least eight times in completely different lines: Hep~'etus, Hydro-
cynus and certain Citharinidae in Africa, Ctenoluciuset aI., Erythrinus et aI.,
Acestrorhynchus et aI. and several other Characinae, not forgetting the
piranhas (Serrasalmidae) in South America.
The plant-eaters belong to at least three families, namely Citharinidae,
Anostomidae and Curimatidae. But some members of other groups, for
example Deuterodon in the Tetragonopterinae, and several others, also
developed a herbivorous habitoThe same is true for the limnivorous species,
found in the Citharinidae as welI as in the Curimatidae, the "Aufwuchs"-
eaters, which nibble food from rocks or falIen trees (two or three families),
the plankton-feeders (at least two families), and the specialized forros
referred to above. Scale-eaters appear in two families, Characidae and
Serrasalmidae, and the fin-biters also in two families, Citharinidae and
(probably) Serrasalmidae.
Several species are adapted to life in river torrents. They belong to three
families, Citharinidae (with Nannocharaxet al.), Characidiidae (with Chara-
cidiumet aI.) and Hemiodidae (with Parodonet al.).
A ventral keel appeared in several subfamilies of the Characidae (Trip-
ortheus, Gilbertolus, Gnathocharax, Pseudocorynopomaand so on), and also
in certain Curimatidae, as welI as in the Serrasalmidae, culminating, in the
Gasteropelecidae, in a complex of adaptations permitting flight a few inches
above the water surface.
FinalIy, sexual dimorphism is evident in the Characidae as welI as the
S. WEITZMAN in GREENWOOD
et a!. (16 families)
(1) CHARACIDAE
(2) ERYTHRINIDAE
(3) CTENOLUCIIDAE
(4) HEPSETIDAE
(5) CYNODONTIDAE
(6) LEBIASINIDAE
(7) PARODONTIDAE
(8) GASTEROPELECIDAE
(9) PROCHILODONTIDAE
14
PRESENT ARRANGEMENT
(14 families)
(2) ALESTIDAE; (6) CRENUCHIDAE;
(7) CHARACIDIIDAE; (13) SERRASAL-
MIDAE; (14) CHARACIDAE
(4) ERYTHRINIDAE
(5) CTENOLUCIIDAE
(1) HEPSETIDAE
RHAPHIODONTINAE, a subfamily of the
CHARACIDAE
(8) LEBIASINIDAE
PARODONTINAE, a subfamily ofthe
HEMIODIDAE
(12) GASTEROPELECIDAE
PROCHILODINAE, a subfamily of the
CURIMA TIDAE
15
~ -- -""
(10) CURIMATIDAE (11) CURIMATIDAE (including
PROCHILODINAE, CHILODINAE and
ANODINAE)
(9) ANOSTOMIDAE
(10) HEMIODIDAE (including
PARODONTINAE)
CHILODINAE, a subfamily of the
CURIMA TIDAE
DISTICHODINAE, a subfamily of the
CITHARINIDAE
(3) CITHARINIDAE (including
DISTICHODINAE and
ICHTHYBORINAE)
ICHTHYBORINAE, a subfamily of the
CITHARINIDAE.
Part I
The African Families
(11) ANOSTOMIDAE
(12) HEMIODONTIDAE
(13) CHILODONTIDAE
(14) DISTICHODONTIDAE
(15) CITHARINIDAE
KEY TO THE AFRICAN FAMILIES*
(16) ICHTHYBORIDAE
a. Scalescycloid; teeth usually well developed, not movable; joint between two sides
of mandible with some interlocking device, movable; no pharyngeal "bags".
b. Predaceous fishes armed with strong canines; maxilla entirely toothed; an
accessory ectopterygoid bone; a large supraopercle bearing a lateral line
canal . . . HEPSETIDAE
bb. Omnivorous fishes usually with incisiform and molariform teeth, rarely
piscivorous and with canines (as in the subfamily Hydrocyninae, with a single
genus); maxilla never toothed; no accessory ectopterygoid bone; no
supraopercle . . . ALESTIDAE
(2 subfamilies)
aa. Scales ctenoid (except in Citharinus, which is a bream-like fish very different
from Hepsetus or from an alestid); teeth usually small, often movable; joint
between two sides of mandible simple, sometimes ankylosed; pharyngeal "bags"
present . . . CITHARINIDAE
(3 subfamilies)
The evolutionary importance of the group will be conspicuous, it is hoped,
in the body of the book, which will deal successively with the African and
American characoids, ending, in a third part, with the most successful
Neotropical group, the Characidae.
Family Hepsetidae
African Pike-like Characoids
Hepsetusodoe,formerly called Sarcodacesodoe,is the single representative
of thefamily. Its external resemblancewithanAcestrorhynchus,byconvergent
evolution, is striking. However, an examination of the finer structures of its
skull reveals a number of differences. Hepsetus, according to Roberts, is
more closely related to another South American group, the Ctenoluciidae,
but with many more primitive features (accessory ectopterygoid bone,
flattened cranial roof, large canal-bearing supraopercle, and so on) which
make it probably the most primitive characoid. The African pike lives in
almost alI rivers of the tropical area, with the exception of the Nile basin.
It is a fish-predator, welI armed with strong, caniniform teeth, similar to
those of Hoplias, Acestrorhynchus and the like. The maxillary bone is
* This key, like ali other keys in the book, is for identificationpurposes only. It is not
intended to reflect relationships.
16
17
entirely toothed as in most predators. According to the author's own
experience in Mrica, it is a good game-fish, behaving like a European pike,
putting up a good fight. It reaches 30 em in length and is edible. Its color is
brownish with indistinct bars on the body and small spots on the fins; its
adipose fin is bright orange and black, apparently forming a recognition
signal.
Svensson, as well as Johnels, has reported that Hepsetus spawns somewhat
in the manner of the anabantids. A nest of floating foam is made by the
spawning pair and the eggs (which are afterwards guarded by both parents
or one parent only) are deposited in it.
Family Alestidae
African Tetras
The Alestidae are the Mrican counterparts of the Characidae, and there
is little doubt that the two groups have the same originoOne example will
show how similar they are. A new genus and species (Pellegrininahetero-
lepis) was once described as a newMrican form dose to Alestesand Brycinus,
because the specimen studied was thought to come from "some part ofWest
Mrica." A more critical evaluation revealed afterward that the label on the
bottle was incorrect and that the fish was a well-known South American
Chalceus.
Despite the general resemblance between the two groups (which led
several ichthyologists to recognize on1yone family from both sides of the
Atlantic Ocean), the Alestidae have the following characteristics in con-
trast to those of the Characidae (at least those characids which most
resemble the alestids, namely the Bryconinae): no rhinosphenoid (a small
bone in front of the orbitosphenoid, present in Brycon); olfactory nerve
endosed in a bony tube formed by the orbitosphenoid and the lateral
ethmoid; lateralline, when present, very low, the number of scales between
lateralline and ventral fin Dever exceeding 3 (usually 2). In the Alestinae,
the maxillary bone (more or less covered by the first infraorbitals) is tooth-
less, usually paddle-like with a rounded contour and a thin pedicel (like a
music note). The premaxillary bone has a characteristic pedicel-likeposterior
process extending to the maxilla, which is rather firmlyunited to it; the inner
premaxillary teeth are invariably four in number (apparently a primitive
number, also to be found in the American Hemibryconini), and molariform
(except in the small species), usually with the crushing face excavated, the
posterior border rounded, multi-cuspidate, and the anterior one with two
prominent cusps or points. On the mandible there is usually a pair of small
conical teeth behind the main series.
18
.-.....
In the Hydrocyninae, which are ichthyophagous predators derived from
some Alestes, the powerful jaws are armed with a single series of very acute
caniniform teeth; it is believed that these teeth correspond to the outer
series of the alestins. Another pecu1iarityofthis group is the type ofmiddle
joint between the two parts of the mandible. It is a "bolted hinge," whereas
that of the alestins is said to be "a fluted tenon." Otherwise, as concems
the general organization, a Hydrocynus does not ditfer' essentially from an
Alestes. Accordingly, the family Alestidae will be split into two groups at
the subfamiliallevel:
a. Teeth in 2 rows on the upper jaw (three rows in one genus), the inner row usually
composed of molariform teethj omnivorous species
. . . ALESTINAE
aa. Teeth in 1 row on both jaws, caniniformj ichthyophagouSspecies
... HYDROCYNINAE
SUBFAMIL Y ALESTINAE
African Tetras
Quite artificially, one may divide the alestins into two groups, Alestini
and Petersini, which could be called tribes. The Petersini looklike the young
of the Alestini and constitute what are properly called the Mrican tetras.
They ditfer from the Alestini not on1y by their small size (frequently in
correlation with the reduction of the lateralline), but also in other "regres-
sive" characters. The inner teeth of the upper jaw are no longer excavated
and molariform, but have retained a single, pluricuspidate cutting edge, and
the inner mandibulary conical teeth are often lacking. Some alestins, the
so-called longipinnis-group,form a SOftoftransition between the two groups,
the species being small and the molariform structure of their teeth scarcely
evident.
Tribe Alestini
The presentauthor recognizes3 generadiagnosedby the followingkey:
a. An adipose lid covering at least part of eye; always a fontanel; scales with
radiating "radii" which are never anastomosed.
b. Premaxillary teeth triserial; 3 scales between lateral line and ventral fin;
dorsal fin well in front of midbody, filamentous in adult male j body deep (depth
2.8-3.5instandardlength) . . . Bryconaethiops
bb. Premaxillary teeth biserialj It-2 scales between latcralline and ventral finj
dorsal fin at midbody or even behind, not filamentous; body elongate (depth
3.25-5instandardlength) . . . A/estes
lia. No adipose lidj usually no fontanel, except in a few species (B. leuciscus,
B. /ongipinnis, B. chapen',B. opisthotaenia) j "radii" of the scales usually
anastomosed .. . Brycinus
. . " ,., "":> Py. Vaniel
I.C''- I 19
C po~rltll.;',- .P \ 8,..,.
~9.000_ Manaus-. Amazona. -
I
GENUS BRYCQNAETHIOPS
This genus includes a few species whieh ehange eonsiderably during
growth, eertain of the adults being seareely identifiable with the young. By
the time they reaeh 200 mm in length they are usually very deep, very dark
(sometimes almost blaek), and the filaments of the dorsal fin of the males
reaeh the caudal finoThe group is at onee reeognized by the 3 series of teeth
on the upper jaw (it is somewhat like modified Alestes, just as Creagrutus,in
the American tetras, looks like a modified Hemibrycon), but is urgently in
need of a revision at the species leveI.
The two or three speeies, B. microstoma,B. boulengeriand perhaps B.
macrops(whieh, if valid, is mueh rarer), are mostly from the Congo basin
(including Ubanghi) although some species are found in Cameroon and
Gaboon.
Alestes macrophthalmusis a southem form from the Congo and Gaboon,
A. liebrechtsiiis restricted to the C~ngo basin, and the small A. stuhlmanni
seems to be an endemic form from East Mrica.
Exeept for some red eolor on the caudal fin in eertain forms, species of
Alestes are usually uniform silvery, without eharacteristie markings.
Bryconaethops sp. Photo by Dr. J. Gery.
GENUS ALESTES
A few species, namely Alestes dentex, A. baremose,A. macrophthalmus,
A. liebrechtsiiand A. stuhlmanni, have rather reeently been separated from
the bulk of the Alestes-like species on the basis of anatomieal details sueh as
adipose lid, fontanels, ete. They are usually large (up to 45 em), elongate
speeies inhabiting large rivers. Some are of eeonomie importanee. Aleste.~
dentex and A. baremose,from the Nile, Niger, Tehad, Senegal, ete., are
sibling species, being separable only by the number of gill-rakers.
A/estes baremose and A. lebrechts (photos by Dr. J. Gery); A.
dentex (after Boulenger, 1909); A. macrophtha/mus, holotype (after
Boulenger, 1909).
20 21
GENUS BRYCINUS
This genus is an aggregate of species in which one may recognize five
groups. Some of these could, perhaps, be ranked as subgenera, although
overlapping of characters is frequent; the others are quite artificial
assemblages. The Brycinus macrolepidotus -group (1), which contains the
type species of the genus Brycinus, B. macrolepidotus,and a few other
species, is characterized by the large size (up to 50em), straight dorsal
profile, dorsal fin far behind mid-body, and overhanging snout. The other
four groups, which have dissimilar characters, may be recognized as
foUows:
a. Interorbital usually less than 3 times in length of head; caudal peduncle high
(its depth 1.0-1.5 in its length).
b. 16-28 gill-rakers on lower arch (fontanel always absent; scales usually with
anastomosed radii) . . . (2) Brycinus nurse-imberi-group
c. Dorsal fin slightlyin front of mid-body . . . B. nurse
cc. Dorsal fin at mid-body or slightly behind . . . B. imberi-group (6-7spp.)
bb. 11-18 gill-rakers on lower arch (fontanel present or absent; scales with or
without anastomosed radiO.
d. Bodydepth3.5-3.66in standardlength; 4 sca1esbetweendorsalfinand
lateral line; sca1eswithout anastomosed radii; no fontanel; dorsal fin
usually behind mid-body . . . (3) B. rutilus-group(2spp. ?)
dd. Body depth 2.3-3.5 in standard length; 5-6 sca1esbetween dorsal fin and
lateralline; sca1eswith anastomosed radii; sometimes a fontanel; dorsal
fin usually at mid-body or in front of mid-body
. . . (4) B.longipinnis-group
(8-9spp.)
aa. Interorbital 3.0-3.5 times in length of head; depth of caudal peduncle 1.5-2.0 in
its length (fontanel always absent; scales without anastomosed radii)
. . . (5) B. humilis-group
(4 spp. ?)
(1) The Brycinus macro/epidotus-group
The large species belonging to the nominal genus Brycinus have a dis-
tinctive shape, the straight snout (which continues the long, also straight,
predorsalline) overhanging the mandible. The smallest of the six induded
species, B. brevis (225 mm) from Nigeria and the Gold Coast, is also the
deepest (body depth 2.75 to 3.0 in standard length), with few lateral line
sca1es(21-23), and is a uniform color, Brycinuscarmesinusis very dose to it.
The four other species are more elongate (body depth 1.5 to 4.0 in the
standard length). Two of them, B. grandisquamisfrom the Congo basin and
B. batesiifrom South Cameroon, attain 26 em in length and can be easily
recognized thanks to a conspicuous spot on the end of the caudal pedunde.
They have, respectively, 23-27 and 28-29 lateral line scales, and 13-14
versus 15-16 anal rays.
22
:"
""" .'
~~ ..4
jj-
Brycinus sp., juvenlle. Photo by G. Timmerman.
Brycinus macrolepidotus, the type species ofthe genus, inhabits a large part
of Mrica from the Nile to the Congo. It usually has a broad dark longi-
tudinal bando Brycinus rhodopleura is its represen~tive in Lake Tanganyika,
differing only in the number of lateralline scales (22-26 sca1es in B. macro-
lepidotus versus 28-29 in B. rhodopleura).They reach, respectively, 42 and
30 em in length.
(2) The Brycinus nurse-imberi-group
Brycinus nurse (type species of the genus BracJzyalestes, not used much
nowadays) and the Brycinus imberi-group (with B. imberi, B. jacksoni,
B. affinis, B. taeniurus, B. kingsleyae and B. bimaculatus), are medium-sized
(100-230 mm) alestins with the dorsal fin at mid-body or thereabouts and
16-28 gill-rakers (usually less than 18 in the Other groups).
Brycinus nurse is the only species in this group to have the dorsal fin slightly
in advance ofmid-body, and it is also the largest (230 mm); it has the same
geographical distribution as B. macrolepidotus, pIus Lakes Victoria and
Rudolf. The caudal fin and part of the ventral and anaI fins are usually red;
humeral and caudal spots are present.
Species ofthe B. imberi-group strictly speaking have the dorsal fin inserted
at mid-body or slightly behind. B. kingsleyae and B. taeniurus from South
Cameroon, Gaboon and Chiloango are immediately recognizable by the
dark lateral band beginning behind the dorsal fin leveI and ending on the
middle caudal rays. Brycinus kingsleyae has a humeral spot (Iacking in
B. taeniurus) and a shorter anaI fin (12-13 branched rays versus 15-17).
23
...
B. imberi (of which curtus, bequaerti and jacksoni are possibly synonyms)
from Cameroon to Kasai, B. bimaculatus from the upper Congo, and B.
affinis, an endemic species from East Mrica, are mostly distinguished by
their color pattern. B. imberi usually has a conspicuous humeral spot and a
broad caudal spot ending on the caudal rays; B. bimaculatus has the same
caudal spot but it is restricted to the caudal pedunde and also has a lateral
spot much farther back (at the leveI of the end of the dorsal fin); B. affinis
has the same humeral spot as B. imberi, but has no caudal spot at all.
elongate body. B. rutilus is a large form (up to 32 em) with the fins some-
times bright vermilion (hence its name). lt inhabits the same regions in West
Africa as B. macrolepidotus,but is nowhere captUred with it (a vicarious
species). B. stolatus from the Congo is a small (?), poorly known species
without color marks.
.
(4) The Brycinus longipinnis-group
This quite unnatUral group forms a sort of transition with the small
alestins called tribe Petersini. Most of the species are small. They some-
times have a fontanel and, at least when young, their inner premaxillary
teeth are more compressed than in typical Brycinus. However, a dose
examination reveals a pair of smalI "shoulders" in front of the largest teeth,
which is almost the only character which distinguishes them from the next
tribe.
B. leuciscus,from the Senegal and Niger basins (where it is calIed tineni
and is of economic importance), has long been confounded with the young
of B. nurse, but it has yelIowfins instead of red. Another species, more
recently described, is alsovery dose. This is B. dagetifrom the Tchad basin,
which differs from B. nurse in being matUre at half the size of the latter
species, as welI as by the number of gilI-rakers. It seems to be the sibling of
B. nurseand might welI be dassified together with B. leuciscus,within the
B. nurse-group.
The B.longipinnis-group strictly speaking is in need ofrevision (like other
Brycinus groups). The folIowing key will aid in distinguishing the species:
a. Not more than 30 lateralline scales (18-21 branched anal rays; dorsal fin filamen-
tous in male; caudal band present) . . . B.longipinnis, B. chaperi
and an undescribed forro
from the Ivory Coast (West
Africa, plus Equatorial
Africa for B. chaperi
on1y ?)
(3) The Brycinus rutilus-group
This small group differs from the preceding one mostly in having fewer
gilI-rakers and from the next one by the structUre of the scales and the more
B;ycinus bimaculatus, hoJotype (after Boulenger, 1899) and B. nurse
(photo by Dr. J. Gery).
aa. 32-37 lateralline scales.
b. 16-21 branched anal rays.
c. An elongate caudal spot; 16-17 branched anal rays; 33 lateralline scales
. . . B. sadleri (endemic species
from Lake Victoria)
cc. Caudal spot broad or absent; 17-21 branched anal rays; 33-37 lateralline
scales.
d. A broad caudal spot; dorsal rays filamentous in male, marked with a black
bar; 33-35 lateral line scales . . . B. intermedius (South
Cameroon)
dd. Body and fins plain; 371ateralline scales. . . B.tessmanni(South
Cameroon)
bb. 24-26 branched anal rays (body and fins plain; 32 lateralline scales)
. . . B. tholloni (Congo basin)
r
24
25
. .
B. longipinnisand B. chaperi,probably sympatric in West Mrica, are the
best known alestins among aquarists. The two species are difficu1tto teU
apart, especially when one has to dea1with young individuaIs. As a role,
B. chaperi has slight1ymore lateralline sca1esthan B. longipinnis(28-30
versus 26-27), as welI as a longer snout, shorter maxilIary bone, broader
interorbital and deeper caudal pedunde. Its dorsal fin is also inserted
farther back than that of B. longipinnis, being at mid-body or slight1y
behind (versus slight1yin front of mid-body).
(5) The Brycinus humi/is-group
The four species. B. humilis, B. abelt~ B. lateralis and B. opisthotaenia are
also very dose to certain species of Micralestes. They are smalI (maximum
size 150mm), with few scales(24-34) and few branched anal rays (12-17).
Brycinus opisthotaenia,from South Cameroon and Gaboon, is the largest
species and has retained many Brycinus macrolepidotuscharacters. It. is
remarkable in having contact tuberdes (similar to those on the snout and
head of mature male barbs and species of Labeo) over the whole body in
both sexes. Nothing is known about the significance of these tuberdes
(calIednuptial tuberdes or pearl organs in the barbs). Like B. opisthotaenia,
B. lateralishas a black lateral band (after preservation) but no humeral spot.
It inhabits Katanga and Angola up to the Zambezi River, and differs by
having more scales(30-33versus 24-27 lateral line scalesin B. opisthotaenia).
B. humilis from Angola and B. abeli from Ubanghi have a horizonta11y
elongate caudal spot; they differ in the same way as the above, having
respectively 30 and 24-26 lateral line sca1es.
Tribe Petersini
5mall African Tetras
This group of about 60 species has recent1y been revised by PolI. He
admitted 16 genera as valid, induding some proposed by Hoedeman in
195.6which are questionable. It looks as if the group is oversplit, which
seems evident when one compares the Petersini with their American counter-
part the Hemibryconini. The latter group has 12 or 13genera with about 80
species. However, the evolutionary conditions were not exact1ythe same,
owing to the absence in South America, for example, of the barbs, which
compete severely with the Mrican tetras. Here on1ya part of the 16 genera
(actually 18 induding Ladigesiaand Virilia, which were not described at
the time of PolI's revision) will be accepted as diagnosed by the following
key:
8rycinus humilis, hOlotype (after Boulenger, 1909) and 8. rutilus
(photo by Dr. J. Gery).
aoTeeth multicuspidate; two premaxillary series.
b. Predorsalline scaled; no outward-directed teetho
c. A pair of conical or (rarely) cuspidate teeth behind the main mandibulary
series.
d. Teeth usually with more than 3 cUSPSo
e. More than 4 outer premaxillary teeth total (lateralline usually complete)
... (I) Micralestes
eeo4 outer premaxillary teeth, 2 on each side (lateral line complete or
incomplete) o o o (2) Phenacogrammus
(This genus maybe split
into 3or 4groups)
dd. Teeth tricuspidate (lateralline incomplete; 4 outer premaxillary teeth)
o. . (3) Tricuspidalestes
cc. Conical teeth behind the main mandibular series usually absento
f. Scales of fiank and abdomen of equal size; 4 or 6 outer premaxillary
teeth (2-3 on each side); 8-11 mandibular teeth.
g. Posterior fontanel present; mandible not prominent (lateral line
completeor incomplete) . . o(4) Hemigrammopetersius
(This genus may be split
into 4 or 5 groups of
species)
26
27
"
gg. Posterior fontanel absent, at least in adults; mandible very pro-
minent (lateralline complete) .. . (5) Petersius
ff. Scales of flank much larger than those of abdomen; 8 outer pre-
maxiUaryteeth; 16-20 mandibular teeth
. . . (6) Arnoldichthys
bb. Predorsalline naked; teeth partly outward-directed
. . . (7) Ladigesia
aa. Teeth conical, in one series (lateralline incomplete).
h. Scalesnormal; mouth cIupeoid
. . . (8) Clupeocharax
hh. Bodynaked,withoutscales;mouthnormal
. . . (9) Lepidarchus
a. Pair of small teeth behind main mandibulary series conical, without cusps; no
more than 7-9 cusps in the broadest teeth.
b. Usually less than 29 lateralline scales.
c. No black longitudinal band; body depth 2.6-2.9 in standard length
. . . M. holargyreus(Congo
basin)
cc. A bIack longitudinal band, more or less complete; body depth 2.9-4 in
standard length.
d. Eye 2.6-3.3 in length of head (adult); more than 50 mm in standard length.
e. 22-24 Iateralline scales.
f. Longitudinal band beginning at dorsallevel; head 3.7-3.9 in standard
length; 15-16branched anal rays . . . M. stormsi(Congobasin;
very difficultto separate
fromM. humilis)
ff. Longitudinal band complete; head 3.95-4.1 in standard length;
16-18branchedanalrays . . . M. vittatus(Lake
Tanganyika)
ee. 25-29 lateralline scales (longitudinal band complete).
g. Lateralline complete.
h. Length of head 3.4-3.8 in standard length.
i. Scales ofIateraI line with bIack spot at base; 25-27 lateralline
scales . . . M.humilis (Tchad, Congo
and Zambezi Rivers)
i i. Scales of lateralline not marked; 24-25 Iateralline scales .
. . . M. voltae (upper Volta
basin; difficult to teU from
M. elongatus)
hh. Length of head 3.75-4.1 in standard length.
j. 18-19 branched anal rays; tip of dorsal fin black (25-26
lateralline scales) . . . M. occidentalis (WestMrica)
j j. 16-17 branched anal rays; tip of dorsal fin hyaline
. . . M. sardina (upper Congo
basin)
gg. Lateral line usuaIIy not quite complete, lacking a few pores on
caudal peduncIe; 16-19 branched anal rays; length of head 3.4-4
in standard length (species of West Mrica).
k. Tip of dorsal fin black; body depth 3.45-3.95 in standard
length . . . M.elongatus (upperVolta)
kk. Tip of dorsal fin hyaline; body depth 2.8-3.3 in standard
length . . . M.comoensis (IvoryCoast)
dd. Eye 2.25-2.5 in length of head (adult); less than 50 mm in standard
length . . . M./odori (Congo)
bb. 29-33 lateralline scales (body depth 3.5-4.2 in standard length; longitudinal
band complete).
1. 16-19 branched anal rays; scales of lateral line not
marked.
m. Caudal peduncIe 1.15-1.4 times longer than deep
. . . M. lualabae(Congobasin)
mm. Caudal peduncIe 1.5-1.75 times longer than deep
M. argyrotaenia(Angola)
The Petersini are thus composed of 6 monotypic genera which are very
well 'Characrerized (Tricuspidalestes, Petersius, Arnoldichthys, Ladigesia,
Clupeocharax and Lepidarchus) and of three polytypic genera, two of which
are quite complicated and perhaps not monophyletic.
(1) GENUS MICRALESTES
Micralestes is the lar,gest genus of the Petersini as well as the most
generalized. All the speciesare of small size(maximumlength 100or 110mm
for the type species, Micralesteshumilis),are moderately elongate, usually
have a longitudinal dark band, and are very similar to each other. According
to Poll, they can be distinguished as follows:
M/era/estes aeut/dens. Photo by Dr. Herbert R. Axelrod.
t",
29
ll. 19-21 h.:r-mchedana1 rays; sca1es of lateral line with
black spot at base (caudal peduncle as deep as long)
. . . M. congicus(Congobasin)
aa. Pair of small teeth behind main mandibulary series with 2-4 cusps; up to 12cusps
in the broadest teeth (longitudinal band complete; tip of dorsal fin black; 14-16
branched ana1rays and 26-29lateral line sca1es) ... M. acutidens (fromNigeria
to Angolaand Zambezi
basin; WestMrica?)
(2) GENUS PHENACOGRAMMUS
This composite genus inc1udes the Micralestes-like species with an
incomplete lateral line (however Micralestes elongatus and comoensis some-
times have an incomplete lateralline), plus certain forms which probably
come from di1ferent evolutionary groups but are not separable at the
generic leveI, their characters showing some overlap. These were considered
as genera by Poll. It seems convenient to retain a similar division, but at a
lower leveI, as follows:
a. Lateralline complete.
b. 4t or 5t transverse sca1esbetween predorsal line and lateralline
. . . Ph. altus-group (or
subgenus Brachypetersius)
bb. 6t-7t transverse sca1esbetween predorsalline and lateralline
. . . Ph. ansorgii-group(or
subgenus Nannopetersius)
aa. Lateral line incomplete (or not quite complete in certain individuaIs) (4t-8t
transverse sca1esbetween predorsalline and lateralline).
c. 15-21 branched anal rays; unpaired fins often filamentous in adult male;
body of moderate depth (2.6-4.3 in standard length)
. . . Ph. interruptus-group
(Phenacogrammus typica1)
cc. 22-25 branched anal rays; unpaired fins unrnodified in adult male; body
usually deep (body depth 1.9-2.9 in standard length)
. . . Ph. caudomaculatus-group
(or subgenus Bathyaethiops,
type speciesgreeni)
The Phenacogrammus altus-group
The small species of this group (Iargest 100 mm) are rather well
characterized by their deep body and relatively small number of transverse
scaIes(four or five between dorsal fin and lateralline), which is unusual (the
number of scale rows theoreticaIly increases with the depth of the body),
as well as by a peculiar spot on the lower part of the caudal pedunc1e,
present in four species out of six (a fifth species has a band going onto the
pedunc1e). These species can be identified as follows:
30
~]
"
,
,I
'
.
' I
fI
11.,
i~
,...
"";~'U~-
!
I
J -- -
Phenacogrammus a/tus, juvenlle. Photo by Dr. Herbert R. Axelrod.
a. 5 scales between dorsal fin and lateral line.
b. A peduncular spot; 16-18 branched anal rays (28-30 lateralline sca1es)
. . . Ph. notospilus(Ogooue,
Gaboon)
bb. No peduncular spot but a broad, diffuse longitudinal band present
. . . Ph. huloti (Congobasin)
aa. 4 sca1esbetween dorsal fin and lateral line.
c. A peduncular spot; 20-27 branched ana1rays.
d. 20-26 lateralline sca1esin a complete series; body very deep, 2.1-2.5 in
standard length (adult).
e. Peduncular spot round . . . Ph. altus, synonym
nummifer (Congo basin)
ee. Peduncular spot horizontally elongate . . . Ph. pseudonummifer(upper
Congo basin)
dd. 25-30 lateralline scales, the lateralline usually lacking one or two pores;
body not very deep, 2.4-2.8 in standard length (adult) (peduncular spot
horiz~ntally elongate) . . . Ph. cadwala~ri(upper
Congo)
cc. No peduncular spot; 18-19 branched anal rays (a prominent humeral spot;
species somewhat intermediary between this and the following group)
. . . Ph.gabonensis(Gaboon)
The Phenacogrammus ansorgii-group
Ph. ansorgiiand Ph. lamberti are small (75 mm) species inhabiting the
coastal part of Gaboon, Congo and Angola. Ph. ansorgii,best known under
the name Petersius ubalo (which was named for the female of ansorgii),is
an elegant fish with strong sexual dimorphism, the males having a con-
31
li. 19-21 !Lranched anal rays; scales of lateral line with
black spot at base (caudal peduncle as deep as long)
. . . M. congicus (Congo basin)
aa. Pair of small teeth behind main mandibu1ary series with 2-4 cusps; up to 12 cusps
in the broadest teeth (longitudinal band complete; tip of dorsal fin black; 14-16
branched anal rays and 26-29lateralline scales) '" M. acutidens (from Nigeria
to Angola and Zambezi
basin; West Arrica?)
(2) GENUS PHENACOGRAMMUS
This composite genus includes the Micralestes-like species with an
incomplete lateral line (however Micralestes elongatus and comoensis some-
times have an incomplete lateralline), plus certain forms which probably
come from different evolutionary groups but are not separable at the
generic leveI, their characters showing some overIap. These were considered
as genera by PoU. It seems convenient to retain a similar division, but at a
lower leveI, as follows:
a. Lateralline complete.
b. 4t or 5t transverse scales between predorsal line and lateralline
. . . Ph. altus-group (or
subgenus Brachypetersius)
bb. 6t-7t transverse scales between predorsalline and lateralline
. . . Ph. ansorgii-group(or
subgenus Nannopetersius)
33. Lateral line incomplete (or not quite complete in certain individuaIs) (4t-8t
transverse scales between predorsal line and lateral line).
c. 15-21 branched anal rays; unpaired fins often filamentous in adult male;
body of moderate depth (2.6-4.3 in standard length)
. . . Ph. interruptus-group
(Phenacogrammus typical)
cc. 22-25 branched anal rays; unpaired fins unmodified in adult male; body
usually deep (body depth 1.9-2.9 in standard length)
. . . Ph. caudomaculatus-group
(or subgenus Bathyaethiops,
type species greeni)
The Phenacogrammus altus-group
The small species of this group (largest 100 mm) are rather well
characterized by their deep body and relatively small number of transverse
sca1es(four or five between dorsal fin and lateralline), which is unusuaI (the
number of scaIe rows theoretically increases with the depth of the body),
as well as by a peculiar spot on the lower part of the caudal peduncle,
present in four species out of six (a fifth species has a band going onto the
peduncle). These species can be identified as follows:
30
iJ!
~
'~1\.'-W~
-- ~- J
Phenacogrammus altus, juvenile. Photo by Dr. Herbert R. Axelrod.
a. 5 scales between dorsal fin and lateral line.
b. A peduncular spot; 16-18 branched anal rays (28-30 lateralline scales)
. . . Ph. notospilus(Ogooue,
Gaboon)
bb. No peduncu1ar spot but a broad, diffuse longitudinal band present
. . . Ph. huloti(Congobasin)
33. 4 scales between dorsal fin and lateralline.
c. A peduncular spot; 20-27 branched anal rays.
d. 20-26 lateralline scales in a complete series; body very deep, 2.1-2.5 in
standard length (adult).
e. Peduncular spot round . . . Ph. altus, synonym
nummifer (Congo basin)
ee. Peduncular spot horizontally elongate . . . Ph.pseudonummifer (upper
Congo basin)
dd. 25-30 lateralline scales, the lateralline usually lacking one or two pores;
body not very deep, 2.4-2.8 in standard length (adult) (peduncular spot
horiz~ntally elongate) . . . Ph.cadwala~ri(upper
Congo)
ec. No peduncular spot; 18-19 branched anal rays (a prominent humeral spot;
species somewhat intermediary between this and the following group)
. . . Ph. gabonensis (Gaboon)
The Phenacogrammus ansorgii-group
Ph. ansorgiiand Ph. lamberti are small (75 mm) species inhabiting the
coastal part of Gaboon, Congo and Angola. Ph. ansorgii,best known under
the name Petersius ubalo (which was named for the female of ansorgii),is
an elegant fish with strong sexual dimorphism, the males having a con-
31
Type of Phenacogrammus ansorgll (after Poli, 1967).
spicuous humeral spot and a rather pale, diffuse lateral band up to the tip
of the middle caudal rays, as welI as filamentous fins, whereas the female
has only the longitudinal color pattem. Ph. lamberti has the same fila-
mentous fins in the males (chiefly the dorsal fin) and a much more con-
spicuous lateral band (from snout to tip of middle caudal rays), as welI as
a characteristic black mark (spot or band) on each caudallobe. Apart from
the coloration, the two species differ in the number of lower gill-rakers,
10-12 in lambertiagainst 14-17 in ansorgii.
The Phenacogrammus interruptus-group
This group is mostly known to aquarists by the type species of the genus,
the Congo tetra, Phenacogrammusinterruptus, which has a particularly
brilliant coloration. Other species, such as Ph. major, Ph. urotaeniaand Ph.
aurantiacus,would be equalIy suitable for the aquarium. One species from
the Congo, Ph. polli, is very different from alI other species, having up to
35 or 37 longtudinal scales and 6 or 7 scales from dorsal fin to lateralline.
It is very smalI (32 mm) with a more or less diffuse lateral bando
A second species, Ph. pabrensisfrom the upper Volta, is also quite dif-
ferent in having a peculiar modification of the anterior lobe of the anal fin
in the adult male; the fiest rays are thickened and curved, the concavity
toward the fore-part. This sexual dimorphism is also to be seen in
Micralestes elongatus, M. brevianalis (a synonym of the West Arrican
populations of Micralestes gcutidens?), and in Hemigrammopetersiusinter-
medius. In other Petersini, mature males have the anterior lobe of the anal
fin thickened, but the rays have their concavity toward the back. This is
why a new genus, Virilia, had been proposed for P. pabrensis. This genus
is not retained here.
32
Hepsetus odoe. Photo by Pierre Brichard.
Hepsetus odoe. Larger individual with spotted fins. Photo by Hllmar
Hansen, Aquarium Berlln.
33
t
'~"." ~ '
1
~
.
'':~
t r!
PhenacogrammusInterruptus. Photo by G. Tlmmerman.
Another group indudes a dubious species (Ph. stigmaturafrom Cameroon)
and 5 very dose forms. Ph. stigmaturahas few longitudinal scales(about 20)
and differs from Ph. interruptusin the presence of an unusual small spot at
the base of the lower caudal lobe.
The last 5 species, which measure about 80 mm in standard length, have
in common their sexual dimorphism, males of most of them having fila-
mentous dorsal and caudal fins, a deeper body and a more brilliant colora-
tion than the females, and many overlapping meristic characters. They falI
nevertheless into two groups, depending upon color pattern: Ph. interruptus
and aurantiacuson the one hand, have a diffuse, often not very conspicuous,
broad lateral band; Ph. major, Ph. urotaeniaand Ph. deheyni on the other
band, have a black band (after preservation) from the level of the dorsal
fin, or farther back, to the end of the caudal rays, this band extending onto
the lower part of the caudal pedunde only.
Ph. interruptus, from the Congo basin (called the Congo tetra by
aquarists) differs from Ph. aurantiacus, also from the Congo basin (and
Gaboon), mostly in the color of their fins, which are mostly pink to purple
in Ph. interruptusand orange (ventral and anal fins) to carmine red (base of
pectoral and dorsal fins and whole adipose and caudal fins) in Ph. aurantiacus.
Both have a body shining with alI the colors of a rainbow. Museum
specimens can be told apart by the number of scales(20-23 in a longitudinal
series, versus 23-26 respectively) and chiefly by the inner mandibular teeth
which are very small or even lacking in 20% of the specimens of Ph.
34
aurantiacus. The latter species is evidently at the limits of the genus
Hemigrammopetersius(where those teeth are always lacking).
The other three forms, Ph. major, Ph. urotaeniaand Ph. deheyni, could
very well be lumped in the future into a single polytypic species. They
usually have a conspicuous humeral spot and the characteristic peduncle
p~ttern described above. The coloration of only Ph. urotaeniais known
with precision. In males the flanks are purple and the belly, the eye and alI
the fins are light carmine. The females are less colorful, more elongate, and
they have two dark lines above their anal fino
The Phenacogrammus caudomaculatus-group
The three Congolese species of the Ph. caudomaculatus-group,or sub-
genus Bathyaethiops, are deep bodied fishes, as deep as Phenacogrammus
altus, which are rather well characterized by a large ovate black spot which
extends (or almost extends) onto the entire depth of the pre-peduncu1ar
region, i.e. just above the last anal rays. They usually have a rather small
humeral spot and a series of chevron-like stripes along the flanks. Ph.
caudomaculatus,the Mrican moon-tetra, is by far tlie most common of the
three species. Ph. greeni (type species of Bathyaethiops) differs from it in
having more transverse, longitudinal and peduncu1ar scales (7!-St/36-42/
41-5.t, 14-16 peduncu1ar, in greeni, versus 5!-7t/28-35/3!-41, 10-12 ped.,
in caudomaculatus).Ph. breuseghemihas the same meristics as Ph. caudo-
maculatus, but it is somewhat less deep, with the peduncu1ar spot more
posterior and only on the lower parto
Phenacogrammuscaudomaculatus (after Poli, 1959).
35
Brycinus macrolepidotus juvenile. Photo by Dr. Herbert R.
Axelrod.
Bryconaethiops sp., apparently not B. boulengeri or B. micro-
stoma. Photo by Dr. Herbert R. Axelrod.
Brycinus imberi from Chete Island cove, Lake Kariba. Photo by Dr. E.
Balon.
r--
I
Brycinus macrolepidotus adult from Gaboon. Photo by Dr. J.
Gery.
~ <;;I-~
Brycinus taeniurus. Photo by Dr. J. Gery.
~
.~
-~
~
~
36
37
d. Broadest premaxillary teeth usual1ywith less than 9 cusps
. . . H. caudalis-group ("genus"
A!estopetersius, rype species
hi!gendorfi)
dd. Broadest premaxillary teeth usual1y with 9-14 cusps
. . . H. tumbensis-group (for
which PolI has proposed
the genus Duboisia!estes)
bb. Lateralline incomplete (except sometimes in H. intermedius); species from the
Cameroon and Tchad . . . H.pulcher-group
(Hemigrammopetersius in the
strict sense)
aa. Preventral region keeled; mouth superior; dorsal fin far behind mid-body
(otherwise very similar to Hemigrammopetersius intermedius)
. . . H. barnardi (East Afric-
probably a new generic
rype)
Trcuspdalestes caeruleus (after Poli, 1967).
(3) GENUS TRICUSPIDALESTES
The single Congolese species, T. caeruleus,has the narrowest teeth of all
Petersini (except Clupeocharaxand Lepidarchus).It is a small fish (44 mm)
having the form of a Micralestesbut with long fin filaments, a large eye and
a large,maxillary bone, somewhat clupeoid in appearance. Its scales are
numerous (36-41 in a longitudinal series, 3-6 of which are perforated, and
7 or 8 from dorsal fin .to lateralline). The overall coloration is steel-blue,
with the head and fins dark, and the back and flanks deep blue or mauve.
It would certainly be a prized aquarium fish.
The Hemigrammopetersius tangensis-group
The H. tangensis-group(sometimes called Rhabdalestes)inc1udes 6 or 7
small (maximum 75~) species mostly from East and South Mrica
(H. tangensis, H. maunensis, H. rhodesiensis and H. leleupi), with a few forms
from West Africa (H. septentrionalis, H. loennbergi and H. eburneensis). They
are usually elongate and resemble certain Micralestes, from which they differ
only, or mainly, in the absence of the pair of small teeth behind the main
mandibulary series.
The East and South African forms may be distinguished as follows:
a. 28-29 longitudinal scales, 10 around caudal peduncle (lateralline complete)
. . . H. maunensis
(Bechuanaland)
aa. 35-37 longitudinal scales, 12-14 around caudal peduncle.
b. Lateralline complete.
c. An inconspicuous dark line on caudal peduncle
. . . H. tangensis (East Africa)
cc. Lateral band conspicuous, mosdy posterior; anal fin base with a black band
. . . H. rhodesiensis(Rhodesia)
bb. Lateralline incomplete, without pores on caudal peduncle; coloration same as
rhodesiensis . . . H. !e!eupi (EastAfrica,
probably an isolatedform
derived fromH. tangensis)
(4) GENUS HEMIGRAMMOPETERSIUS
Like Phenacogrammus, Hemigrammopetersius has been split into four
subgroups which have generic status according to Poll. However, the
characters alleged for the division show considerable overlap and the
genera cannot be retained here. This overlap (with the exception of a
specialized form, H. barnardi) is clearly apparent in the following key:
a. Preventral region not keeled; mouth terminal or not much superior; dorsal fin not
much behind mid-body.
b. Lateral line complete or not quite complete, rarely incomplete.
c. Elongate specieswithout filaments on the fins, derived fromMiera!estes,with
usual1y more than 2 outer premaxillary teeth on each side; 2 subgroups,
East African and West African . . . H.tangensis-group ("genus"
Rhabda!estes)
cc. Deeper bodied species with filamentous fins on the males, resembling
Phenaeogrammusand perhaps derived from it, with 2 outer premaxillary
teeth; al1species Congolese except one from Nigeria.
The West African forms are even closer; all have 14-17 branched anal
rays, and the numbers of scales and gill-rakers overlap. They also have in
c0!llmon a black longitudinal band, more conspicuous posteriorly, and a
black anal fin margin. The only differential characters are as follows:
38
39
i
I
~ ".~
-~
,
'\
#
~
"
Brycinus chaperi. Adult male. Photo by Or. Herbert R. Axelrod.
Male (above) and female (below) Brycinus lateralis from Chikanka
Island cove, Lake Kariba. Photo by Or. E. Balon.
" ~
~ 't'lf
~";.JoA'" (j
.4""'y ..,.
~:;;:
~/
....
',,"
" "'~
,
-
J'
~)'
f'
-
/'
Young Brycinus chaperi from Lagos. Photo by Or. Herbert R. Axelrod.
Brycinus longipinnis, male above and female below. Photo by J. Elias.
a.Tip of dorsal fin black.
b. Lateralline complete, with 22-26 tubes (sometimes with a pair of conica1teeth
behind the main mandibu1ary series; a connecting link with Micralestes ?)
. . . H. eburneensis(endemicin
the Cavallyin Ivory Coast)
bb. Lateralline perforating only 17-24sca1es . . . H. loennbergi (Gambia,
Comoe in Ivory Coast, and
BlackVolta)
aa. Tip of dorsal fin hya1ine . . . H. septentrionalis (from
Senegal to Cameroon)
One of the types of Hem;grammopeters;us tangens;s (after Poli,
1967).
The Hemigrammopetersius caudalis-group
The seven species of the H. caudalis-group constitute a rather homo-
geneous group, with the exception of H. hilgendorfi (type of the genus
Alestopetersius),which is quite aberrant in having 34-42 lateral line scales
(instead of 27-33 for the bulk of the group) and up to 8 scales betWeenthe
dorsal fin and lateralline. Six of the seven species are Congolese and the
males of 6 out of 7 species have a filamentous dorsal fin, while 2 have the
middle caudal rays prolonged as in certain species of Phenacogrammus.
The species are relatively easy to identify, as shown in the followingkey.
The commonest species are H. hilgendorfiand H. caudalis,which is some-
times found in the aquarium trade under the name Petersiusxenurus.
42
I
Key to the species of the H. caudalis-group:
a. 27-33longitudinal scales (plus 1or 2 on caudal fin); 5-7 transverse scales between
dorsal fin and lateralline.
b. 13-20 gill-rakers on lower part of first arch.
c. Fins hyaline.
d. No broad lateral band from eye to end of middle caudal rays (often a
lateral streak which may get broader on caudal peduncle).
e. Caudal fin of males without prolonged median rays.
f. Body depth 3.0-3.15 in standard length; 27 + 2 longitudinal scales;
16 gill-rakers; a conspicuous triangular caudal spot; humeral spot
scarcelyvisible . . . H.smykalai (Lower Niger)
ff. Body depth 2.4-3.0 in standard length; 30-33 + 1 or 2 longitudinal
sca1es;13-14gill-rakers; no caudal spot; humeral spot large, vertica1ly
elongate . . . H.compressus (Congo)
ee. Caudal fin of males prolonged (body depth 2.5-2.8 in standard length;
27-32 + 1 or 2 longitudinal scales; 13-18 gill-rakers; humeral spot
vertically elongate; a black lateralline extending onto caudal rays, where
it is much broader) . . . H.caudalis, synonym
H. xenurus(Congo)
dd. A broad lateral band, equally broad from eye.to end of prolonged middle
caudal rays (body depth 3.25-4.0 in standard length)
. . . H. brichardi (Congo)
Hem;grammopeters;us caudalis. Photo by Milan Chvojka.
43
Mcralestes storms. Photo by Or. Herbert R. Axelrod.
Mcralestes acutdens from Northern Rhodesia. Photo by Or.
Herbert R. Axelrod.
44
,..........--
Young Phenacogrammus nterruptus. Photo by Or. Herbert R. Axel-
rodo
Male Phenacogrammus nterruptus, the Congo tetra. Photo by. Or.
Herbert R. Axelrod.
,..
~ -"- -~
~~. ~
~
.~
.~ ~
" ~
~
~~
:
,
45
cc. Fins black, particularly the filamentous dorsal fin and the caudal fin (except
at its base) (middle caudal rays not prolonged; a humeral spot; body depth
2.9-3.4 in standard length; a sibling species of H. hilgendorfi ?)
. . . H. nigropterus(Congo)
bb. 24-29 gill-rakers on lower part of mst arch (dorsal and caudal fins not pro-
longed; a broad lateral band from eye to end of middle caudal rays; dorsal,
pectoral and ventral fins black) . . . H. leopoldianus (Congo)
aa. 34-42 longitudinal scales(plus 1or 2 on caudl fin); 7-8 transverse scales between
dorsal fin and lateralline (dorsal fin filamentous, middle caudal rays not pro-
longed; a vertica1lyelongate humeral spot; fins black in adult male)
. . H. hilgendorfi(Congo)
The Hemigrammopetersius tumbensis-group
This group, for which PoU proposed a new generic name (Duboisialestes)
owing to its extremely broad teeth (6 to 14 cusps instead ofusually less than
9 cusps for the H. caudalis-group), is composed of only 2 species, H. tum-
bensis from the Stanley Pool (Lower Congo), known as H. duboisi by
aquarists, and a much rarer species, H. bifasciatus, from the Middle Congo
basin. The latter has more scales (7t/34-36/3t, 13-15 ped., in H. bifasciatus,
versus 5t-6t/30-33/3t, 10-12 ped., in H. tumbensis) and conspicuous black
marks on the base of the caudal lobes. H. tumbensis resembles H. caudalis,
with which it has been confused having been described as a subspecies of
H. xenurus (which is a synonym of caudalis).
Anal fln of a female (above) and male Hemigrammopeterslus inter-
medlus (after Gery, 1968).
46
.I-
Hemigrammopetersius pulcher, type (after Poli, 1967) and H. tumben-
sis, type (after Poli, 1967).
The Hemigrammopetersius pulcher-group
One species surely belongs here, the poor1y-knownH. pulcher ftom the
Ia River in Cameroon, because it is the type of the genus Hemigrammo-
petersius. Only two specimens are known, which look similar to the Phena-
cogrammusmajor-urotaenia-deheynicomplex, but without the pair of small
conical teeth behind the main mandibulary series.
Three additional species are provisionally referred to the 'group by PoU,
H. brevidorsalisand H. intermediusfrom the Tchad and Chari River, and
H. barnardifrom East Africa. The first two species are said to differ mostly
in the number of gill-rakers. However, H. brevidorsalisis known on1yfrom
damaged adult specimens and some juveniles. Males of H. intermediusare
rather similar to those of H. septentrionalis,and certain specimens even have
the latera1line complete or almost complete. They have a thickened, rounded
anal fin, surrounded by a broad black bando
H. barnardiis different from aUother species of the group, being adapted
to a life near the surface. It may be a new generic type, but for the moment
is retained in Hemigrammopetersius.
47
Phenacogrammus aurantiacus. A male individual without very much
color in its fins. Photo by Dr. Herbert R. Axelrod.
Male Phenacogrammus urotaenia. Photo by Dr. J. Gery.
Phenacogrammus ansorgii. Photo by Dr. Herbert R. Axelrod.
Phenacogrammus huloti. Photo by Pierre Brichard.
--.......
-:;..,.-
;...
48
49
(5) GENUS PETERSIUS
The single species Petersius conserialis,from around Dar-es-Salaam in
Tanganyika (coast of East Mrica), is quite unusuaI among this group of
genera in attaining a relatively large size (145mm), the mouth uptumed
with the mandible prominent, and in Iacking a posterior fontaneI. It has a
caudal spot extending onto the middIe caudal rays. This is about alI that is
known of this species.
(6) GENUS ARNOLDICHTHYS
The single(?) species Arnoldichthysspilopterus,the Niger tetra or Arnold's
characin, is well known to aquarists. It is unusual not only in having dif-
ferent sizes of scales, but also in having a larger number of teeth than other
Petersini. The "role" of 4 inner premaxillary teeth, for example, which
has fewexceptions, is not valid here. The Niger tetra has up to 7 inner teeth
on a side. The same is ttue for the lower row of teeth and, in fact, lacking
the pair of conical teeth on the mandible, this Mrican tetra has almost the
same dentition as that of its American cousins. However, the shape of its
body, and chiefly of its snout, is distinctly alestid, indeed, much more
Brycinus-like than Micralestes-like.
Its rainbow coloration is well known and need not be described here.
The sexual dimorphism of its anal fin is less well known. In the female, the
anal fin, with the border almost straight like that of other alestins, has a
black spot in the Iast rays; in the male the anal fin, with a convex edge, is
tricolor, red and yellowwith a black border. According to certain aquarists,
there is a polymorphism in this color pattem, some males (or males of a
sibling species?) having an irregularly striated anaI fino
Petersius conserlalis and Clupeocharax schoutedenl (aftar Poli, 1967).
(7) GENUS LADIGESIA
A single speciesfrom Sierra Leone, has recent1ybeen described, Ladigesia
roloffi, but slight1ydifferent forms have been collected since in other parts
of West Mrica.
The jelly bean tetra is another peculiarity of the Mrican tetras. It has a
good part of the predorsal region entirely naked, and its upper teeth point
outward. Its body shape and black color pattem resemble that of an
"ordinary" Hemigrammopetersiussuch as H: septentrionalisor H. inter-
medius, but the adult males (largest size 35-40 mm) are entirely bright
vermilion.
retained the body shape of a Petersini, but has acquired a clupeoid, uptumed
mouth with a blade-like maxillary in a remarkable convergency with some
Aphyoditeini (American), Brittanichthys for example. This smaIl fish
(47 mm) is a species from the region ofLake Tumba (near the confluence of
the Ubanghi with the Congo), where it performs seasonal migrations,
according to Matthes, as well as from another lake of the central Congo. It
is a pelagic plankton-eater and, as such, is not very colorful, being without
marks except for a narrow lateral line.
(8) GENUS CLUPEOCHARAX (synonym CLUPEOPETERSIUSj
The single species Clupeocharax schoutedeni (do not confuse it with
Clupeacharax,which is a South American genus) is again aberrant: it has
(9) GENUS LEPIDARCHUS
The single species Lepidarchus adonis if. one of the most striking dis-
coveries of recent years. This minute (less than 20 mm standard length)
West Mrican fish is entirely devoid of scaIesand lateralline. The teeth, like
those of Clupeocharax,are in a single series on both jaws, min~te, conicaI,
but the maxillary is not clupeoid.
The male has numerous purple spots on the distal haIf of its body as well
as on the caudal finoThe ventral and anal fins are likewise spotted or even
barred, whereas the front part of the dorsal fin is brown in both sexes.
50
51
~! ~
'1
....
Hemigrammopetersius caudalis, the yellow-tailed Congo tetra.
Photo by Dr. Herbert R. Axelrod.
Hemigrammopetersius intermedius. Photo by H.J. Franke.
----
Phenacogrammus caudomaculatus, the African moon-tetra.
Photo by Dr. Herbert R. Axelrod.
Hemigrammopetersius rhodesiensis. Photo by Warren E. Bur-
gesso .
52
53
~
SUBFAMILY HYDROCYNINAE
African Tiger Fishes
Hydrocynus (not Hydrocyonas it is frequently spelled) species are A/estes
modified toward predation. They have acquired strong, acute, uniserial
canine teeth, those of the upper jaw altemating with those of the lower and
acting as a trapo This makes the tiger fishes some of the most formidable
fish-killers of African waters.
Six species are known, difi"eringin the number of scales and in certain
proportions. H. forskalii and H. vittatus, very dose species, are widely
distributed in Mrica. They are found from the Nile, Senegal and Niger
to the Congo, with H. 'Ottatuseven extending to South Mrica. H. brevis
and H. somonorum,siblings in Niger and Tchad, are restricted to West
Mrica, whereas H. vittiger and the giant tiger fish, H. goliath, inhabit the
Congo. Most of the species reach 40 or 50 em, or even 68 cm for H. brevis.
The record size is held by H. goliath, which attains a length of 132em and
a weight of 38 kg, and which is said to reach 150em and 50 kg on occasion.
Certain species are brilliantly colored during the breedin~ season.
H. vittatus for example, the most ubiquitous, haS a series of longitudinal
lines formed by spots at the base of each scale, has the eye and ventral fins
red, and has the other fins at least partially orange.
Most of these characters are quite specialized. The three subfamilies
can be told apart by means of the following key:
Key to subfamilies of Citharinidae:
a. Premaxillary bones short and not movable; joint between two sides of mandible
movable.
b. Sca1escyc10id(except in one species); body short and compressed, bream-like
. . . CITHARININAE
bb. Scales always ctenoid; body short and compressed only in a few instances
(XenocharaxandDistichodus) ... DISTICHODINAE
aa. Premaxillary bones long and movable; joint between two sides of mandible not
movable . . . ICHTHYBORINAE
Family Citharinidae
Citharinids
SUBFAMILY CITHARININAE
Citharins
This smalI group of large, usually deep-bodied, limnivorous species is
exactly homologous to the American prochilodins and, like the latter, are
of economic importance in certain regions of tropical Mrica. The largest
species reaches a size of 80 em and a weight of 18kg. The scales are cydoid
in Citharinusand ctenoid in Citharidium, and are numerous (46 to 90 in the
lateralline, which is complete). The dorsal and anal fins are relatively long
(16-24 rays for the former and 19-31for the latter) in relation to the great
depth of the body. The mouth structures are well adapted for feeding on
the mud at the bottom of rivers: it is very broad, with very weakteeth inser-
ted into the lips; the maxilIary bones are not toothed; the gill-rakers are
numerous; and the intestinal tract is very longo
The citharinids, strictly Mrican, are sometimes split into 2 families
(Distichodidae and Citharinidae) or even 3 (with the Ichthyboridae). These
groups will be treated here as subfamilies which have the following
characters in common:
-Scales ctenoid (with the exception of the genus Citharinus).
-Ventral rays usualIymore numerous than in the other characoids(except
hemiodids).
-Only two epurals.
- Teeth small, often movable, usually cuspidate (again a convergence
with the hemiodids).
-Presence of pharyngea1"bags".
-According to Daget, interca1arium very small and incorporated into
the neurocranium (whereas in a non-citharinid such as Brycon, it is a
separate small osseous plate covering the prootic and other bones of
the cranium).
-Joint between the two sides ofthe mandible simple, without interIock-
ing devices, often ankylosed.
GENUS CITHARINUS
Seven species belonging to the genus Citharinus (which is split by some
into three subgenera) are known. The folIowing key wilI help in the identi-
fication of these species (after J. Daget):
a. Relatively generalized forms with body not very deep and adipose fin base not
very long (shorter than its distance to the dorsal fin).
b. 7 or 8 transverse sca1erows between lateralline and ventral fin; 46-51 lateral
line sca1es; body depth 2.2-2.7 in standard length
. . . C. distichodoides (Tchad and
Niger basin)
bb. 10-20 transverse scale rows between lateral line and ventral fin; 55-90
lateral line scales; body depth usually less than 2.3 in standard length.
c. 10-12transverse scale rows between lateralline and ventral fin; 55-60 lateral
line scales . .. C.macrolepis (Congo basin)
cc. More than 12transversescalerows; more than 60 lateralline sca1es.
d. 13or 14transversescalerowsbetweenlateralline and ventralfino
e. 60-661ateralline scales . . . C. congicus (Congobasin)
ee. 68-73lateralline scales . . . C.eburneensis (IvoryCoast)
..,; ~
td~/IJ Mele,,! Rap1' PII. Da,det
C. po.,." ~,:j-',\'I',\ "
89.000- M.na.. - AmazoA" - a~
55
54 "
...............
"
I
,
I
f
/
~
Unidentlfied Afrlcan tetra. Photo by Klaus Paysan.
I'
..
.
\'
\\
Arnoldichthys spilopterus, tlTe Niger tetra. Photo by Dr. Herbert R.
Axelrod.
Ladigesia roloffi, the jelly bean tetra. Photo by E. Roloff.
-
Lepidarchus adonis. Photo by E. Roloff.
.-...
56
57
dd. 17-20 transverse scale rows between lateralline and ventral fin (77-90
lateralline sca1es) . . . C.citharus, typespecies of
the genus Citharinus(from
Senegal to Nile basin)
aa. Specialized, deep-bodied forms with an elongate adipose fin base (longer than its
distance from dorsal fin).
f. 10-12 transverse scale rows between lateral line and ventral fin;
53-60 lateralline sca1es(same sca1ationas that or C. macrolepis)
. . . C. gibbosus(Congobasin)
ff. 13-15transverse sca1erows between lateralline and ventral fin; 59-71
laterallinescales . . . C. latus(fromSenegalto
Nilebasins)
GENUS CITHARIDIUM
The second genus, Citharidium, with the single species ansorgii,from the
Niger basin, is exactly like the deep-bodied Citharinus with a long adipose
fin, but its scales (46-50 in lateralline and 11or 12 between lateralline and
ventral fin) bear longitudinal (alongaxis ofbody), parallel ridges terminating
in spines. According to Daget, these sca1esare different from the typical
ctenoid scales of the next subfamily.
SUBFAMILY DISTICHODINAE
Distichodins
The Distichodinae is a relatively large group of about 50 or 55 micro-
predator and herbivorous species. They are distributed among fewer than
10 genera, as follows:
a. Generalized forms with gill membranes not rormmg a continuous bridge under
throat; body deep, head broad and snout obtUse (two or three rows of teeth on
both jaws) . . . (1) Xenocharax
aa. More specialized forms with gill membranes forming a continuous bridge under
throat; body shape variable (two series of teeth on both jaws).
b. Gill membrane not fused with throat or isthmus; maxillary teeth often present
(small rather elongatespecieswith head narrow and snout more or less pointed).
c. 31-48 lateralline (or longitudinal) scales.
d. Lateralline complete; adipose fin present
(2) Nannaethiops
dd. Lateral line incomplete; adipose fin present or absent.
e. 10-14 dorsal fin rays; maxillary teeth present
.. . (3) Neolebias
ee. 15-16 dorsal fin rays; maxillary teeth present or absent.
f. A series of pit-lines on head; maxilla toothless; adipose fin present
. . . (4) Congocharax
fr. No pit-lines on head (?); maxilla toothed; adipose fin absent
. . . (5) Dundocharax
cc. 59-73 lateral line sca1es; a genus forming the transition with the
Ichthyborinae . . . (6) Microstomatichthyoborus
58
.,......--
PI xxxn
f,
....
i;,or
Type of Citharinus gibbosus (after Boulenger, 1899) and type of Cltha-
r/d/um ansorgii (after Boulenger, 1909).
59
--
Hydrocynus sp., possibly H. goliath or may be a giant specimen of H.
vttatus. Photo by Dr. J.P. Gosse.
60
. .'..- -- "-
~ji'. fJ ' -
:7
~\.
,', ~ iL ",.' . v
,,,"" .,.,' ("~ '"
I . . . I I ' ' ,/" .1 "1.;'. '- "-:;"",
. , ' 'r
. . . ,>~,. .." . ... ~(
~
~
.~...
..
'>i..-
..,., -".:-
1.-."
1
, J.
""""
Hydrocynus vttatus juveniles trom Lake Karlba. Photo by Dr. E.
Balon.
Ctharnus ctharus. Photo by Klaus Paysan.
bb. Gill membrane fused with isthmus; maxilla toothless.
g. Two series of bifid teeth on both jaws; dorsal fin long, with 16-27
rays.
h. Large deep bodied species, quite comparable to the Citharl-
ninae; adipose and caudal fins sca1ed
. . . (7) Distic1wdus
hh. Small,elongatespecies;adiposefin not sca1ed
... (S)Paraduc1wdus
gg. A single series of teeth on both jaws; dorsal fin relatively short,
with 12-16 rays.
i. Lateralline complete
i i. Latera1line incomplete
. . . (9) Nannccharax
. . . (10) Hemigrammocharax
relatively large size: 62 mm) . . . genus Nannaethiops (a
single species, N.
unitaeniatus, from Niger to
the Congo basin)
aa. Lateral line incomplete (adipose fin present or absent; inner teeth bifid or
conical on both jaws; 2-7 maxillary teeth) . . . genus Neolebias
b. 16 scales around caudal pedunc1e; a single longitUdinal bando
c. About 7 maxillary teeth; a broad dark band along the side and a dark bar at
base of caudal fin; adipose fin usually absent
. . . Neolebias ansorgii (from
Cameroon-called
N. landgrafi-to lower
Congo basin)
cc. Alwaysless than 5 maxillary teeth; a rather narrow band along the side and
no dark bar at base of caudal fino
d. A longitUdinal black band extending to caudal fino
e. LongitUdinal black band extending along middle of caudal pedunc1e;
usually a caudal spot; maxillary bone not reaching leveI of eye.
f. 33-35 longitUdinal sca1es; body depth 3.0-3.7 in standard-Iength;
dorsal fin behind mid-body; usually no adipose fin
.. . Neolebias unifasciatus (type
of the genus, Portuguese
Guinea to Gaboon)
ff. Usually 36-39 longitudinal scales; body depth 3.9-4.15 in standard
length; dorsal fin in front of mid-body; adipose fin present
. . . N eolebias gracilis (central
Congo basin)
ee. LongitUdinal black band extending along lower part of caudal pedunc1e;
no caudal spot; maxillary bone almost reaching to leveI of pupil;
(adipose fin present; 33-36 longitUdinal sca1es; body depth 2.5-3.3 in
standard length; dorsal fin in front of mid-body)
. . . Neolebias trewavasae (Nile
to lower Congo)
dd. A longitUdinal band interrupted above anal fin, not continued on caudal
pedunc1e(no adipose fin; maxillary bone not reaching leveI of eye; dorsal
fin at mid-body; 31-33 longitUdinal scales)
. . . Neolebias axelrodi (a species
described from 2 small
specimens less than 19 mm
in standard length collected
. near Lagos, Nigeria, by
Dr. H. R. Axelrod)
bb. 12 or 13 sca1esaround caudal pedunc1e; 2 or 3 longitudinal bands.
g. Two longitUdinal bands; no caudal spot; no adipose fin
. . . Neolebias philippei (Central
Congo basin)
gg. Three longitUdinal bands; a small caudal spot; adipose fui usual1y
present . . . Neolebias trilineatus (Congo
basin); often called
Nannaethiopstritaeniatus
(1) GENUS XENOCHARAX
Xenocharax spilurus,ca1ledIhajo by the Kota fishermen in Gaboon, is a
deep bodied citharinid which differs from the other large species (of
Citharinus, Citharidiumand Disticlwdus)only by a few details such as the
structure of the gill membranes (which are separated in the midd1eas in
most characoids). It reaches 25 em in size and, chiefly when young, is quite
nicely colored with the fins orange or red. It inhabits Cameroon, Gaboon
and the Cotigo.
A second species, Xenocharaxcrassus,found in the Upper Congo basin, is
said to differ from the preceding mostly in some proportions and in the
presence of a black spot on the dorsal fino
(2-5) NANNAETHIOPS AND ALLlED GENERA
This group of 10 or 11 pigmy distichodins comprises the genera
Nannaethiops and Neolebias,very popular among aquarists, as well as two
other more recent (and much less known) genera from the Congo basin
and Angola, Congocharaxand Dundocharax. Dur,docharaxwas previously
c1assifiednear Hemigrammocharax,with which it does not seem to have
much affinity.
The first two genera have the following characteristics: small size;
maxillary bone always toothed; two rows of teeth on both jaws, the outer
ones bifid, the inner ones also bifid (N. unitaeniatus, N. trilineatus, N.
trewavasaeand maybe others) or conical; 11-15 dorsal rays; 31-36 longi-
tudinal scales. The following key is inspired by the rather recent revjsion
of Poll & Gosse (1963). The subgenera proposed by Daget the following
year are not used.
Key to the genera Nannaethiops and Neolebias:
a. Lateralline complote (adiposefin present; inner teeth bifid on both jaws; usually
2 maxillaryteeth; a longitUdinaldark band extending onto middle caudal fin rays;
62
63
J
D
Type of Nannaethiops unitaeniatus (after Guenther, 1871).
All species are micropredators, living on smaIl animais found on the
bottom or among aquatic plants. The foIlowing species are easy to acclimate
to the aquarium andoare often seen. (I) N. ansorgii and a paler form ca1led
N. landgrafi, wbich probably belong to the same species. The former variety
is green-blue (a quite unusual tint among fishes) with red fins; the latter
form is pale blue. (2) N. trilineatus, wbich has its median longitudinal band
surrounded by a golden metallic stripe like that of Rasbora pauciperforata or
Hemigrammus erythrozonus, as weIl as reddish finnage. (3) N. unifasciatus,
wbich is said to have the same color pattern but without the upper and
lower dark bands. (4) N. trewavasae, wbich was often confused with the last,
though it is somewhat less elongate and with the fins not strongly colored.
GENERA (4) CONGOCHARAX AND (5) DUNDOCHARAX
The genus Congocharaxwas erected for a species previously c1assified
within lfemigrammocharax (olbrechtsi) as well as for Neolebiasspilotaenia.
The type species has the maxillary bone not toothed, it has slightly more
dorsal rays and more scales than the Neolebiasspecies, and has some pit-
lines on the head. The two species are distinguished as follows:
a. 40-43 longitudinal scales; a lateral band) wIch may be broken into a series of
spots . .. Congocharaxspilotaenia
(Chiloango basin)
aa. 44-48 longitudinal sca1es;a humeral spot and a large caudal spot
. .. Congocharax olbrechtsi
(Central Congo basin)
66
With regard to Dundocharax, the single species D. bidentatus (from
Angola) differs from the preceding species by having 30-32 longitudinal
scales, some maxillary teeth and no adiposefinoJudging from its description,
the genus is scarcely separable from Neolebias.
(6)GENUS MICROSTOMATICHTHYOBORUS
Tbis long, difficult name designates quite interesting Congolese species
somewhat intermediate between the distichodins and the ichthyborins.
The body and head are elongate, the teeth much like those of a Neolebias,
the sca1esnumerous (59-73 in the lateralline, which is complete), and the
caudal fin is obliquely marked like that of a scissortail rasbora. The two
known species are micropredators like Neolebias, and not ichthyophagous
or fin-eaters like most ichthyborins. They can be distinguished as foIlows:
a.59-61 lateralline sca1es
aa. 63-73lateralline scales
. . . M. katangae
. . . M. bashfordeani
.
Microstomatichthyoborus bashfordeani (after Fowler, 1936).
(7) GENUS DISTICHODUS
Here we havea series of large (maximum size 70 em) detritivorous and
herbivorous species which differ from a preceding genus' Citharinus,
mostly in having the sca1esctenoid, the teeth biserial and the adipose fin
as weIl as the caudal fin lobes covered with smaIl sca1es;the caudaIlobes are
frequently rounded. Boulenger, in bis c1assica1Caralogueof the Fresh-
water Fishes of Africa, listed 17 species of Distichodus.Later, a few more
species were described, and the foIlowingkey wiIl help to characterize and
identify the species:
67
fi
a. Less than 60 lateralline scales (16-20 dorsal rays; snout broad or feebly com-
pressed).
b. Ana! fin with 19-22 rays, at least as long as dorsal fino
c. 7/37-39/10 scales . . . D. affinis(Congobasin)
cc.9/40-42/11-12 scales . . . D. altus(upperCongo)
bb. Anal fin with 12-17 rays, shorter than dorsal fino
d. Mouth terminal; dorsal fin with a black blotch or bar; 38-46 lateralline
scales.
e. 12 or 13 anal rays; very small
(6-7/38-41/6-8 scales)
n. 15-16 scales from mid-dorsal line to lateral line
(13-15 scales from lateralline to base of ventral
fin) . . . D.rostratus (Nile, Senegal
and Niger)
nn. 18-20 scales from mid-dorsal line to lateral line
(15-18 scales from lateralline to base of ventral
fin) . . . D. niloticus(Nileonly)
mm. Body spotted, even in adults (17-20 scales from
mid-dorsalline to lateralline; 18-20 from lateral
line to base of ventral fin)
. . . D. brevipinnis (Nile, Senegal
and Niger)
size; a series of spots on body
. . . D. decemmaculatus (a pigmy
species from the Central
Congo basin)
ee. 14-16 anal rays; moderate size; body plain.
f. 6 or 7 scales from lateral line to base of ventral fin; caudal lobes
pointed . . . D. notospilus (fromSouth
Cameroun to Angola)
ff. 8 scales from lateralline to base of ventral fin; caudallobes rounded
. . . D. noboli (upper Congo)
dd. Mouth strongly inferior; dorsal fin plain; 53-571ateralline scales (12or 13
anal rays; 6 scales from lateralline to base of ventral fin; caudallobes
obtusely pointed) . . . D. hypostomatus (Gaboon)
88. 60-110 lateralline scales.
g. Snout deeper than longo
h. Snout broad or feebly compressed (mouth terminal to strongly
inferior).
i. Mouth terminal or only slightly inferior.
j. 60-78 lateralline scales.
k. Snout shorter than interorbital space, 1.33-1.66 in its
width (dorsal fin spotted)
1. 9 or 10 scales betweW1lateral line and base of ventral
fin; 60-66lateralline scales; about 14narrow transverse
bars on body . . . D.antonii (Congo basin)
11. 13 or 14 scales between lateral line and base of ventral
fin; 68-75 lateralline scales; a humeral spot
. . . D. petersii (East Mrica)
kk. Snout about equal to or slightly shorter than interorbital
space (lI or 12 scales between lateral line and base of
ventral fin; 67-70 lateralline scales)
. . . D. mossambicus(Zambezi
River*)
j j. 80-110 lateral line scales (snout much shorter than inter-
orbital space; mouth sub-inferior; large species from the
Nile and Niger, verycloseto each other, the dorsal fin always
more or less spotted, the body spotted in the young, usually
plain in adult (except brevipinnis).
m. Body plain (adults).
ii. Mouth strongly inferior.
O. 16-19 dorsal rays; 11 anal rays (series of spots
on flank; 9-10/68-75/10-12 scales)
. . . D. maculatus (Congo basin)
00.22-27 dorsal rays; 13-16 anal rays.
p. Snout shorter than interorbital space,
1.33-1.66 in its width; (6 transverse bars on
body; 13-14/68-77/14-15 scales)
. . . D. atroventralis (Congo
basin)
pp. Snout about equal to or slightly shorter than
interorbital width.
q. Body plain (71 lateralline scales; 12 bet-
ween lateral line and base of ventral fin;
very dose to D. mossambicuswith the
exception of the position of the mouth)
. . . D. langi (Congo basin)
qq. Body pattemed.
r. About 18 transverse bars on body
(13-15/66-78/14-16scales)
. . . D. fasciolatus (Congo basin
and Angola)
rr. Series of spots on body (15-16/75-85/
16-17 scales)
. . . D. engycephalus (Nile and
Niger basins)
hh. Snout strongly compressed (mouth sub-inferior; 24-25 dorsal
rays, 12-14 anal rays; 14-16/60-68/14-16 scales; 6 or 7 broad
transverse bars on body) . . . D.sexfasciatus (Congobasin
and Angola)
gg. Snout longer than deep (mouth more or less terminal; 25-26 dorsal
rays, 13-14 anal rays; 15-17/70-85/15-17 scales; 6 to 8 broad
transverse bars on body) . . . D. lusosso (Congobasinand
Angola)
* The individuais mentioned in the literature as having the mouth inferior may
belong to D. langi.
The author does not possess enough data concerning Distichodusalbini
from East Mrica (apparently near D. petersii), D. kolleri(synon~ vexillifer)
70
71
...
Young Distichodus notospilus. Photo by Or. J. Gery.
A juvenile Distichodus, possibly D. nobo/i. Photo by Or. Her-
bert R. Axelrod.
Distichodus nobo/i. Photo by Or. Herbert R. Axelroa.
Adult Distichodus notospilus. Photo by Or. J. Gery.
72
73
from Cameroon (dose to D. notospilus), and D. schenga (dose to mossam-
bicus?) to fit them into the key.
(8) GENUS PARADISTICHODUS
Paradistichodusdiffers from Distichodusby the more elongate body and
the adipose fin not being covered with small scales. The single species,
Paradistichodus dimidiatus, rather strongly resembles a Nannocharax. It
differs from Nannocharaxby its biserial teeth, as well as by having a longer
dorsal fin than in most Nannocharax species (formula iii or iv, 13-15). It
has a complete lateralline, with 51-63 perforated scales, and can at once be
recognized due to its characteristic dorsal spot (which may act in nature
as a recognition signal). Some ichthyologists acknowledge two subspecies,
the nominal one from Casamance to the upper Niger basin, and a second
one from Tchad.
'1if
Paradstchodus dmdatus. Photo by Dr. J. Gery.
GENERA (9) NANNOCHARAX AND
(10) HEMIGRAMMOCHARAX
AFRICAN DARTERS
Nannocharax and Hemigrammocharaxare the exact counterpart in Mrica
of Characidium(and allied genera) of South Arnerica. Both strongly con-
vergent groups are bonom fishes with large pectoral fins and sometimes
other adaptations to life in fast-running waters, even torrents. A specimen
of Nannocharaxis usually impossible to distinguish from Characidiumwith
the naked eye (provided that the collecting locality is not given), except
for a certain rugosity of the body surface due to the ctenoid scales, which
contrasts with the smoothness of the cydoid scales of the Characidium.
As contrasted with the other distichodins, Nannocharax and Hemigram-
mocharax have a single series of bifid teeth on both jaws. Like their
Neotropical cousins, they are rather poorly known, as far as their taxonomy
is concerned, primarily because most of the original descriptions were in-
complete. This is why the following key will indicate only the outlines of
their dassification:
74
r
a. Lateralline complete; mouth usually inferior; body usually somewhat compressed
. . . genus Nannocharax*.
b. Dorsal fin inserted in front of levei of ventral fins (usually less than 50 lateral
line scales except in N. luapulae).
c. 14 branched dorsal rays; body depth 3.5 in standard length
. . . N. ogoensis (Gaboon)
cc. 9-12 branched dorsal rays; body depth more than 3.5 in standard length.
d. Dorsal fin inserted at mid-body (even more posteriorly in young)
. . . N. pareus (Niger to Ogowe;
possible synonyms are
ansorgei and micros)
dd. Dorsal fin inserted in front of mid-body.
e. Less than 46 lateralline scales; body depth 4.5 in standard length.
f. About 35-38 lateralline scales . . . N. pteron (Ubanghi)
ff. About 40-44 lateralline scales . . . N. macropterus (Congo,
Angola)
ee. 48-50 lateral line scales; body depth 4.5-6.5 in standard length.
g. 6 or 6t transverse scales belowlateralline; 6 or 7 round &potsalong
body . . . N.gracilis (Congo)
gg. 7t transverse scales below lateralline; about 10 vertically elongate
spots along body . . . N. taenia andN. luapulae
(Ubanghi to South Katanga)
bb. Dorsal fin inserted at levei of ventral fins or behind it, the ventral fins usually
being very anterior (37-57 lateralline scales).
h. 37-42 lateralline scales.
i. 6 branched anal rays (broad spots on the dorsum)
. . . N. brevis(upper Congo
basin)
i i. 7 or 8 branched anal rays.
j. Depth usually less than 4.33
. . . N. altus and N. ho//yi
(Ubanghi and Gaboon)
j j. Depth 4.25-5.0 . . . N. procatopusand
N. schoutedeni(Congo to
Angola)
hh. 42-57 lateralline scales.
k. Body depth usually less than 5.5 in standard length;
several vertical bars on body.
1. 4t scales above lateralline; 42-49 lateralline scales
. . . N. fasciatus, type of the
genus (with a subspecies in
the,J.iberian-Guinean forest
and the nominate fonn from
Niger to Gaboon)
11.5t scales above lateralline; 47-55 lateralline scales
. . . N. intermedius(South
Cameroon)
* The author lacks data conceming N. gobioidesRoman from the Upper Volta.
75
A young Distichodus
sexfasciatus. Photo
by B. Kahl.
Distichodus fasciolatus. Photo by Dr. Herbert R. Axelrod.
76
Distichodus sexfasciatus adult. Photo by Dr. Herbert R. Axelrod.
77
kk. Body depth usually more than 5.5 in standard length;
body marked with roundish spots.
m. 1unbranched ray at beginning of ventral fin
. . . N. lineomaculatus(Niger
and Tchad)
mm. 2 unbranched rays at beginning of ventral fin
. . . N. niloticus (with 4
subspecies from the Nile,
nominal form, to the
Congo basin)
aa. Lateral line usual1y incomplete (said to be complete in certain individuaIs of
H. multifasciatus); mouth usual1y terminal or subinferior; body usual1y quite
cylindriform (an ocellus at root of caudal fin) ...genus Hemigrammocharax
n. 34-44 longitUdinal scales.
o. 12 scales around caudal pedunc1e.
p. Adipose fin always (?) present.
q. 170r 18 dorsal rays and 15 or 16 anal rays
(35 longitUdinal scales)
. . . H. uniocellatus(type
species of the dubious
subgenus Microdistichodus;
Congo basin)
qq. 12-16 dorsal rays and 10-12 anal rays
(34-44 longitUdinal scales)
r. Dorsal fin inserted in front of leveI of
ventral fins, or just at its leveI.
s. 21-23 lateral line scales; length of
pectoral fin 1.15 in length of head;
7-10 irregular transverse'bands
. . . H. angolensis(Angola)
ss. 10 lateralline scales; length of pectora1
fin 1.8 in length of head; 14 transverse
bars ... H. monardi(ZambeziRiver)
rr. Dorsal fin inserted behind leveI ofventral
fins (a longitUdinal band, crossed by 9-12
short vertical bars)
. . . H. lineostriatus(Angola)
pp. Adipose fin always (?) absent (13 or 14dorsal
rays and 12 anal rays; dorsal fin behind
ventral leveI; 34-36 longitUdinal scales, of
which only 6-7 are perforated)
. . . H. machadoi(Angola)
00. 15-16 scales around caudal pedunc1e (adipose
fin always (?) present; dorsal fin inserted in
front of leveI of ventral fins; 13-15 dorsal rays
and 10-12 anal rays; body with 13-20 more or
less regular vertical bars).
t. Body depth 3.7-4.7; length of head
3.4-3.8 in standard length; usually
12-32 lateralline scales.
78
.
~
f'
''c'
t
u. Rarely more than 18 lateral line
scales
. . . H. witte; (Katanga;
Angola)
uu. Usually 27-32 lateralline scales
. . . H. stigmaturus-complex
with H. stigmaturus(Zambezi
River), H. minutus (South
Katanga) and H. polli
(upper Volta basin)
n. Body depth up to 5.5; length of head
4.3 in standard length; lateral line
very long or even complete (?) in
some individuaIs
. . . H. multifasciatus(a species
from the Zambezi River,
which is often synonymized
with H. stigmaturus)
nn. About 50 longitUdinal scales
H. ocellicauda(type of the
genus; Cameroon)
Hemigrammocharax wittei (after Poli, 1967) and Nannocharax cf. fas-
ciatus (photo by Dr. J. Gery).
~
-
,.",
....
~.
.
.
'
I
..
.J.
\
'.
'
...
'\\~ ~;i~ .
:I~i1. .
'
."",:. ~t;~'
.. .' ~
/~t IIT
Distichodus sexfasciatus, young fish. Photo by Dr. Herbert R. Axel-
rodo
Distichodus lusosso, young fish. Photo by Dr. Herbert R. Axelrod.
'.'
(:
../ .
..
.
1
: .
.
J
.,... .
~- .
r
~...-
~i
Distichodus ~A'J{fR~r.iRtll.r::_ adulL Photo bv G. Meola. ATrlCan Fish
Imports.
Distichodus lusosso, adult. Photo by Dr. Herbert R. Axelrod.
SUBFAMILV ICHTHVBORINAE
Fin-eaters
This small group of highly specialized predators is very interesting from
a biological point of view, some of its members having a semi-parasitic life,
living, at least in part, by eating the fins of other species. Indeed, with their
scissor-like jaws (the upper one movable upwards), their cutting teeth and
pike-like body, species of the genera Eugnathichthys, Belonophago,Ichthy-
borus and Phago (and possibly others) are well adapted to this specialized
predation, whereas other genera such as Phagoborusand Mesoborus eat
whole fishes, as their mouth, armed with some canine-like teeth, indicates.
Finally, the least specialized of ali in some respects, Hemistichodus,forming
a sort of transition with the preceding subfamily (for example with
Microstomatichthyoborus),can only capture small insects with its bifid
lateral teeth. It should be noted that most of the species have an extremely
small maxillary bone, in contrast with other predators.
The genera can be distinguished as follows: Hemistichodus lootensi and H. mesmaekersi (after Poli and Daget,
1968).
a. No teeth on the front pan of the upper jaw (a single sedes of teeth on the sides)
(I) Hemistichodus
(1) GENUS HEMISTICHODUS
Two or, more probably, three small (50-100 mm) species belong to this
genus. H. vaillanti, the type species from Gaboon, has a characteristic
color pattern: the dorsal and caudal fins are marked with conspicuous black
spots surrounded with orange-yellow; the spot on the lower caudallobe is
the largest, extending onto the distal half of the lobe. It is interesting to note
that almost ali Ichthyborinae have some caudal pattern, usually one or two
bands crossing the lobes obliquely. It has been hypothesized that these
markings act as a recognition signal, telling the other individuais not to
attack their own species.
The species, with two of them sympatric in the central Congo, can be
distinguished by means of the following key:
a. No marks on caudallobes; last unbranched dorsal ray longer than head (64-74
longitudinal scales; lateral line interrupted near middle)
. . . H.lootensi (Central Congo)
aa. Caudallobes with a pattern; last unbranched dorsal ray not longer than head.
b. About 65 longitudinal scales, lateralline interrupted near middle; two hori-
zontal bars on dorsal fin, twOoblique bars on each caudal lobe
. . . H. mesmaekersi(Central
Congo)
bb. About 75-80 longitudinal scales, lateralline complete; a single spot on dorsal
fin and a single spot on each caudal lobe (lower one very large)
. . . H. vaillanti (Gaboon)
aa. Front teeth present, lobed, conical or canine-like.
b. A single series of teeth on both jaws.
c. 2 canines above and 3 below in front; snout not longer than postocular part
ofhead . . . (2) Phagoborus
cc. 8 or 9 canines above and below; snout more than 2 times longer than
postocular part of head . . . (3) Gavialocharax
bb. Two rows of teeth on both jaws.
d. Inner teeth consisting of a pad of brush-like, very small teeth; outer teeth
ca{liniform in front, bifid on sides . . . (4) Ichthyborus
dd. Inner teeth in a row, not brush-like.
e. 2 canines above and below . . . (5) Mesoborus
ee. No canines (fin-eaters for the most part).
f. Snout not longer than postocular part of head.
g. Snout rounded, jaws massive, outer bicuspid teeth all of the same
size; cheek partly naked; gill membranes free from isthmus
. . . (6) Eugnathichthys
gg. Snout narrow, beak-like, outer upper bicuspid teeth larger in front
than on sides; cheek almost entirely covered by suborbitals; gill
membranes united to isthmus.
h. 87-90 lateralline scales, 10 between dorsal fin and lateralline
. . . (7) Paraphago
hh. Less than 50 lateral line scales, I or 2 between dorsal fin and
lateralline . . . (8) Phago
ff. Snout longer than postocular part of head
. . . (9) Belonophago
82
83
""-
,..
--
..,
"
1
.
!
Juvenile Distichodus schenga (above) and D. mossambicus from
Namazambwe. Photo by Or. E. Balon.
Distichodus hypostomatiJs. Photo by Or. J. Gery.
Adult Distichodus schenga. Photo by Or. E. Balon. Adult Distichodus mossambicus. Photo by Or. E. Balon.
(2) GENUS PHAGOBORUS
The type species, Phagoborusornatus (formerly Neoborus ornatus), in-
habits the Central Congo basin as do most Ichthyborinae. It attains a length
of up to 200mm and is a predator of small fishes. It has three conspicuous
black lateral bands on its body, and its orange caudallobes are barred with
three longitudinal bands, more regular on the upper lobe.
A second, smaller (?) species, P. quadrilineatus,with the same caudal
pattem, has been described from a very remote river basin, that of the
Casamanza in Portuguese Guinea. It has fewer lateralline scales(90 instead
of 98-110), fewer anal rays (15 instead of 17-18), and 4 longitudinal bands
on the body.
(3) GENUS GAV/ALOCHARAX
The single species Gavialocharax monodi has been found in Cameroon.
It has the same size and body shape as that of Phagoborus,but it is more
elongate with the snout strongly resembling that of an American ctenolucid.
It has only 2 longitudinal bands on its body, and a very similar caudal fin
pattem.
Ichthyborus besse; dentition and caudal fin pattern (after Poli 1957).
86
~
Type of Gavialocharax monodi (after Pellegrin, 1928).
(4) GENUS ICHTHYBORUS*
The single species, Ichthyborus besse, is remarkable in having evolved
(according to Daget) differentIy in the Nile, Tchad and Benou basins on
the one hand, and in the Congo basin on the other. In the Tchad basin, for
example, it is a fin-eater, at least partIy because there is no other fish to
occupy this very peculiar ecological niche. In the Congo basin, where the
fin-eaters are well establishedand well adapted, the subspeciesl. besse
congolensiscontents itself with small whole fishes and some crustaceans.
The species reaches 200 mm in size and the two forms can be distin-
guished by their caudal fin color pattem. In the nominalform (l. besse
besse)the yellowcaudal fin is spotted. In the formfromthe Katanga(I.
bessecongolensis) these spots are united in vemcular,irregularlines and
there is an ocellus at the fin base.
.1
I
I
I
I
I
(5) GENUS MESOBORUS
The two Mesoborusspecies are apparentIy purely ichthyophagous and
are not well known, being rare. The type species, Mesoboruscrocodilus,from
the Upper Congo, is a rather large fish(250mm long) decorated with 2 series
of round spots on the body and plain fins (or perhaps faded). The second
species, M. pellegrini, also from the Congo basin but from different rivers
(in South Cameroon and Kasai), has more branched dorsal rays (15 or 16
instead of 14) and the dorsal fin and caudal 10bes are decorated with 3
oblique dark bands. The body patterns are the same, and in its general
shape and coloration a Mesoboruslooks much like an Eugnathichthysor a
Paraphago.
* Not Ichthyoborus, which is the generic name of certain birds.
87
Distichodus
decem maculatus.
Photo by B. Kahl.
Distichodus decemmaculatus. Photo by A. Norman.
88
~
Nannocharax fasciatus. Photo by Hilmar Hansen, Aquarium Berlin.
Hemigrammocharax sp., probably H. multifasciatus. Photo by Warren
E. Burgess.
89
I
1
~-::
;::';.
...
"
n
~
Mesoborus pellegrlnl (after Stelndachner, 1913, as M. crocodllus);
Eugnathlchthyseetveldll(after Boulenger, 1898).
(6) GENUS EUGNATHICHTHYS
The two sympatric Eugnathichthysspecies from the Congo basin, both
fin-eaters, are the ecological counterparts of the above-described Mesoborus
species (ichthyophagous), in the same way as are the two subspecies of
Ichthyborusbesse(1. bessebesseis a fin-eater, I. bessecongolensisan ichthyo-
phage). In contrast with Mesoborus,Eugnathichthyshas no caniniform teeth
but has bicuspid, compressed teeth forming a continuous, jaw-like cutting
tool adapted to the fin-eating habit, as was shown by Matthes in the English
journal Nature in 1961. Both species have the dorsal fin and caudallobes
barred with 3 (4?) black bands, and a spotted body.
Eugnathichthys eetveldii, the type species, attains a rather large size
(30 em) and has more scales (95-107 lateral line scales) than the smaller
E. macroterolepis(68-80 lateralline scales), and the spots on the sides are
roundish instead of being vertica1lyelongate.
90
~
)
1
1
(7) GENUS PARAPHAGO
Technically, the single species Paraphago rostratus, from the Congo,
differs from Phagoonly in its numerous sca1es,which is a rather poor generic
character. However, a look at its body shape and coloration (the so-called
habitus ofthe species) shows that it is much more like an elongate Mesoborus
(body depth 5.5-6 in the standard length) that has lost its canine teeth and
has become a fin-eater, with the same teeth as those of Phago.
l
,.
..
?Phago boulengerl. Photo by Tlmmerman.
(8) GENUS PHAGO
The three or four species of Phago ate relatively small (110-160mm),
and have the teeth compressed and tricuspidate, very trenchant and set
dose together (about 17-22 outer teeth on each side of each jaw), which
enables them to cut the fins of other, sometimes larger, fishes. The resem-
blance of their "beak" to scissors is increased by the peculiar upper jawof
the Ichthyborinae, which moves upward when the mandible goes down.
Another peculiarity is the scales, which are rugose, with a median spine
forming a sort of longitudinal crest, giving somewhat the aspect of the
osseous plates of a seahorse.
The species have a fusiform body shape, their fins are barred as in most
other genera, and their elongate body is usually spotted.
The following key gives sufficient data for their identification:
a. "Beak" shorter than postocularpart of head; 42-45laterallinescales,8around
caudalpedunc1e(bodydepth 6-7 in standardlength)
. . . P. boulengeri (Central and
upper Congobasin;
Tchad ?)
91
............
Hem;stlchodus vail/ant;. Photo by Dr. J. Gery.
Paraphago rosttatus. Photo by Dr. Herbert R. Axelrod.
.,
;i'"
~
,
,
J
I
Two characterlstlc habitats of the African characolds. Above: Pool in
Slerra Leone. Photo by E. Roloff. Below: Trlbutary of the Congo Rlver.
Photo by Dr. Herbert R. Axelrod.
aa. "Beak" longer than postocular part of head; 47-48 lateralline sca1es,6 around
caudal peduncle.
b. 2 scales between dorsal fin and lateralline; body depth about 6 in standard
length . . . P.intermedius (Congobasin)
bb. 1 scale betWeendorsal fin and lateral-line; body depth 8-9 in standard length.
c. 19outer teeth on eachside . . . P.loricatus, typespecies
(Niger)
cc. 22outer teeth on eachside . . . P.maculatus (adubious
species, also from the Niger)
---
~
"'1I':""'''I!If!>:..",
Belonophago tinanti. Photo by Tlmmerman.
(9) GENUS BELONOPHAGO
The species of this genus are very unusual in a group where odd forms
are numerous. They resemble a needlefish (family Belonidae) as the generic
name implies, with their very elongate body (body depth 10-11 times in
the standard length) and snout (which is 1.5 to 2 times longer than the
postocular part of the head). The scales, with their median spine, are similar
to those of Phago, as is the scale formula (30-35 in the lateral line, one
between dorsal fin and lateralline, and 6 around the caudal peduncle), and
the fish is something like a very elongate Phago, with the same fin-eating
habits. As concerns the coloration, the usual black marks on the caudal fin
are replaced by red spots (one on each caudal lobe base).
The two species, both from the Congo basin, can be separated as follows:
a. "Beak" twice as long as postocular part ofhead ... B. hutsebouti (Central
Congo basin)
aa. "Beak" only 1.5 to 1.75 times longer than postocular part of head
. . . B. tinanti (StanleyPooI)
94
)
,
\,
\
Part II
The American
Characoids other than
the Characidae
KEV TO THE AMERICAN FAMILIES
a. Teeth, when present, in a singleseries at least on upper jaw; anal fin short, usually
with less than 3 unbranched rays (except in Chilodinae) and less than 10branched
rays (except in Abramites and a few species in the Curimatidae); scales usually
of the alestid-type, i.e. with their circuli of the caudal (apical) part toward the
median line or even concentric (except in the Characidiidae).
b. Amia- or pike-like predators; teeth usually conical; occipital crest truncate;
fontanels lacking.
c. Amia-like predators; dorsal long in Hoplias; caudal fin rounded; maxilla
long; scales large, cyc10id . . . ERYTHRINIDAE
cc. Pike-like predators; dorsal short, inserted far bacI< (together with anal fin);
caudal fin with well formed lobes; maxilla short; scales small, ctenoid
. .. CTENOLUCIIDAE
bb. Usually non-predaceous fishes of medium to small size; teeth conical only in
small or pigmy species (Characidiidae for example); occipital crest and
fontanels variable.
d. Poeciloid body shape; dorsal fin long; sexual dimorphism present; a
paired frontal organ (anterior fontanellacking)
. . . . CRENUCHIDAE
dd. Body-shape variable; dorsal fin short (except in Elachocharax);usually
no external differences between sexes (except in Curimatopsis);no paired
frontal organ.
e. Teeth numerous, often in two rows on mandible; pterygoid teeth often
present; fontanels usually absent (body usually elongate).
f. Scales of the Hemigrammus-type; dorsal fin with up to 19 rays;
usually strongly adapted to life on river bottoms, with a large pectoral
fin, flattened ventral surface, inferior mouth, etc.; maxilla usually
not toothed .. . CHARACIDIIDAE
(3 subfamilies)
ff. Scales of the Alestes-type; dorsal fin ShOIt; species adapted to life at
the surface or in mid-water, with small pectoral fins, rounded ventral
area, terminal or upturned mouth, etc.
LEBIASINIDAE (2 or 3
subfamilies)
95
1
{
,
"
Erythrinus erythrinus. Photo by Harald Schultz.
The trahira, Hoplias ma/abaricus, from the Rio Aguaro. Photo by Dr.
Herbert R. Axelrod.
-- - -- .
},
11'
~'
96
Hoplerythrinus unitaeniatus. Photo by Harald Schultz.
Young Hoplias ma/abaricus. Photo by Dr. Herbert R. Axelrod.
1:1
11
97
ee. Teeth, when present, few in number, with a single series on mandible;
no pterygoid teeth; fontanels usually present; maxilla not toothed
(except in Parodontinae); body shape variable.
g. Teeth always present in adult, at least on upper jaw, and more or
less firmly set into the bone (except in Bivibranchia et a!.).
h. Not more than 4 teeth, strongly tixed into the bone on each jaw
(on each side); maxilla never toothed; mouth usually terminal
or upturned, rarely inferior . . . ANOSTOMIDAE(2 subfamilies)
hh. Usually more than 4 compressed teeth, not firmly fixed into
premaxilla; maxilla and mandible toothed only in the Paro-
dontinae; mouth usua11ysubinferior or inferior
. . . HEMIODIDAE (2-3
subfamilies)
gg. Teeth usually absent, or, when present, set in the lips
. . . CURIMATIDAE
4 subfamilies)
aa. Teeth variable, usually in more than one row on upper jaw (exceptions occur
in the Characinae, Cheirodontinae and Serrasalmidae); anal fin moderate or long,
with at least 3 unbranched rays and 10 branched ones (with a few exceptions in
certain regressed Tetragonopterinae and Cheirodontinae); scales usually cycloid
(except in certain Characinae) with circuli of caudal (apical) zone parallel or even
divergent with axis of body (called Hemigrammus-type, with exceptions in
specialized groups such as the Serrasalmidae and certain Characidae).
L Coracoids forming a large, rounded keel; pectoral fins well
developed, permitting flight (body very compressed; dorsal
fin with up to 17 rays) . . . GASTEROPELECIDAE
(2 subfamilies)
iL Coracoids and pectoral fins usually not developed; when
developed, not being adapted to real flight.
j. Maxilla reduced, not toothed; dorsal fin long, with at least
16 rays; usually a predorsal spine and a series of ventral
spines (serrae) (body very compressed, usually disciform;
teeth variable according to diet; scales small, the circuli
concentric) . . . SERRASALMIDAE
(3 subfamilies)
j j. Maxilla usually toothed and not reduced; dorsal fin short;
rarely a predorsal spine (in the Stethaprioninae); no real
serrae (a strong keel with pointed scales in Stichonodon)
. . . CHARACIDAE (12 subfamilies)
~
Head of a glant trahira (Hoplias macrophthalmus) about 50 em long.
.Photo by Dr. Herbert R. Axelrod. .
which is becoming rare),* but interesting from the zoologist's point ofview.
Some are predators which have become adapted to life in small brooks or
ponds, where they can exist without much oxygen, thanks to accessory
systems which can absorb it directly from the air. At least one species is
capable of retaining moisture and air for a long time and walking on the
ground at night in the search for a more suitable pond than the one just
abandoned.
As regards their anatomy, particularly that of the bones of the head,
certain experts believe that they are exceptional among the characoids.
They are probably the most archaic in many respects (though specialized
for predation), having 5 branchiostegal rays, a cylindrical body form
with a short anal fin and large scales, numerous teeth on the palate and
(apparently) a primitive pectoral girdle like that of the bowfin.
The trahiras may be divided into two groups at the generic leveI, as
follows:
Family Erythrinidae
Trahiras
The trahiras, as the Brazilians call them, are quite similar to the bowfin
(Amia calva) of sluggish waters of the eastern United States, although
they comefromavery different evolutionary line. It is a small group, without
economic importance (except for the giant trahira, a most delicious fish,
a. Dorsal rays iii, 11to 15; at least 37 lateralline scales; maxillary bone with 2 or 3
small canines plus a series of conical teeth . . . Genus Hoplias
aa. Dorsal rays iii, 8 or 9; 32-37 lateralline scales; maxillary bone without canines
Genera Erythrintis and
Hoplerythrinus
*At least in the Guianas.
98
99
Headof Erythrinus erythrinus. Upper Juruena, Mato Grosso. Photo by
Harald Schultz.
Boulengerella Iateristriga. Photo by Dr. J. Gery.
100
-- -
Head of Ctenolucius hujeta. Photo by Dr. Herbert R. Axelrod.
Ctenolucius hujeta. Photo by Dr. Herbert R. Axelrod.
101
GENUS HOPLlAS
There are probably only 3 species in this genus, induding the true
trahira (Boplias malabaricus),which is distributed in most parts of South
America and is somewhat confused as concems its nomendature. It is one
ofthe most often cited ofthe characoids (more than 80 references), but often
under quite different names (see artide by Dr. Viggo Schultz in Tropical
Pish Bobbyist, September, 1971).
The three species can be distinguished by means of the followingkey:
a. Large size(up to one meter); eye relativelylarge, 14 or 15times in standard length
of adults, partly visible from above; abdomen flat and not colored
. . . H. macrophthalmus
(Guianas, where it is called
Aimara)
aa. Rather small size (up to 25 em or so); eye smaU,about 18to 20 times in standard
length of adults, lateral, not visible from above; abdomen round and colored
b. 43-47lateral line s'Cales;ii, 13-15dorsal rays ... H. microlepis(Pacificslope
of southem Ecuador,
Panama and Costa Rica)
bb. Usually 37-43 lateralline scales; iii, li or 12 dorsal rays, rarely 13
. . . H. malabaricus(South
America, including the
countriesbetWeenthe i<
localities of H. microlepis)
*H. lacertae, classicaUy considereI! as a synonym of the trahira (H. malabaricus),
may be a distinct species.
,
Erythrlnus erythrlnus In typlcal surroundlngs. Upper Juruena Rlver.
Photo by Harald Schultz.
----
,;
.~
~.
fi/!
;.
102
GENERA ERYTHRINUS AND HOPLERYTHRINUS
The two genera, each containing a single species, are very dose to each
other. The following key summarizes their most important features:
a.MaxiUary bone not very elongate, not reaching posterior margin of eye in adults;
eye very smaU, 22-25 times in standard length; dorsal fin pointed and elongate in
males; caudal fin sponed; color panern of flanksrather variable, usuaUymarbled
. . . Erythrinus erythrinus
(South Americaand
Trinidad)
aa. MaxiUarybone elongate, its tip weUbeyond posterior margin of eye in adults; eye
smaU, about 19 times in standard length; dorsal fin low in both sexes; caudal fin
plain; dark longitudinal band, caudal spot and opercular oceUuspresent, which
are characteristic ofthe species . . . . Hoplerythrinus unitaeniatus
(South America)
Family Ctenoluciidae
South American Pike-characoids
')
The Ctenoluciidae are medium to large size predators (20-100 cm),
having acquired a very long and tapered pike-like body, with a gavial-like
snout and short dorsal and ana1 fins inserted posteriorly. The lower jaw,
slight1y less elongate than the upper one, is induded; the numerous teeth
are small, usually acute, dose together and directed backward, and occut on
both jaws as well as on the ectopterygoid bone; either the snout or the chin
has fleshy appendages. According to T. Roberts, who studied their anatomy
extensively, they bear the dosest resemblance to the African Hepsetidae.
Both families are rather remote from the AcestrorhynclJjni, the most
elongate members of which have acquired the same pike-like aspect by
evolutionary convergence.
The four known species can be distinguished by means of the following
key:
a.Premaxillae joining in front in a semicircle; fleshyflaps present on chin; a short,
internal second row of teeth on the mandible; ectopterygoid bone (in roof of
mouth) with conical teeth only; lateral line incomplete; 42-50 strongly ctenoid
scales in a longitudinal row; anal fin before middle of dorsal fin, its last rays never
filamentous (an oceUuson caudal fin base) . . . Genus Ctenolucius(Panama
fII11dN.W.South America)
(a single species, C. hujeta,
with 3 subspecies: C. hujeta
beani in Panama and
Colombia, C. hujeta
insculptus in Colombia, and
C. hujeta hujeta in
Venezuela; synonym
Luciocharax insc}41ptus)
103
Head of Boulengerella maculata. Photo by Harald Schultz.
Boulengerella maculata. Photo by Harald Schultz.
--
...,
I
104
\-
Crenuchus spilurus male. Photo by Harald Schultz.
A pair of Poecilocharax weitzmani. Photo by E. Roloff.
105
Photo by Or. Herbert R. Axelrod.
Photo by Or. Herbert R. Axelrod.
Photo by Or. Fernandez-Yepez.
108
'It
These photos represent tive dltterent specles 01 the genus Chara-
cidium. For accurate identlflcations at the species levei an anatomical
study ot the specimens wouid be necessary. Top photo on this page
by Or. Herbert R. Axelrod, bottom photo by Stanisiav Frank.
n"
Poecilocharax bova/lll, male above, female below (after Eigenmann,
1912).
Finally, the most peculiar feature of the Crenuchidae is a mysterious
organ' on top of their head, which is housed in a pair of recesses in the
frontal bone at eye levei and which is not possessed, as far as known, by any
other teleost group. The organ is formed by rod-shaped cells (similar to
those of the retina) encircled by a net of capillary blood vessels. The whole
structure suggests a receptor ofradiations (infrared??) which are compared
with some internal parameter (perhaps the internal temperature?), thanks
to the blood vessels. As the physiological study of this phenomena has not
even begun, the above is pure speculation and a fascinating problem
remains to be solved.
GENUS CRENUCHUS
The single known species, Crenuchusspilurus, from the Guianas and a
large part of the Amazon basin (including the Upper Solimoes and Rio
Negro), differs from the two species of Poecilocharaxmostly in having an
adipose finoThe adult male reaches less than 50 mm in standard length,
and resembles certain dwarf cichlids, due to its body shape as well as to its
brilliant coloration.
110
r
~
It has long been considered an aggressivefish and has never been popular
in aquaria (hence not collected). Therefore, little is known about its
reproductive behavior. Some have said that it spawns on large leaves or
stones (again like certain cichlids). With the great popularlty of cichlids, it
is hoped that this very unusual characoid will become a favorite of
aquarists.
GENUSPOECUOCHARAX
The type of the genus, Poecilocharaxbovallii, is a small, nicely colored,
killy-like fish inhabiting, exclusiveiy, the surroundings of the famous
Kaieteur Falls in Guyana on the Potaro River. According to its first
collector and describer, the famous Carl Eigenmann, the male has a
geranium-red caudal fin decorated with pearl-white spots. As earlyas 1954
Eigenmann's pupil and follower, G. S. Myers, recommended it as a potential
ornamental aquarium fish.
The second species, P. weitzmani, more recently discovered, is even more
striking. It inhabits the Upper Solimoes region (habitat of the well-known
neon tetra), as well as the Upper Negro-Orinoco regions. Adult males of
less than 30 mm standard length have a yellowbody, darkening above and
tending toward red on its posterior part, with a longitudinal black band on
which shines (like a liner in the night with ali portholes illuminated) the
neon-blue center of each scale; superimposed on it is a blood-red band;
the large fins are yellowwith brown and red, as well as pearl-white spots;
the eye is green and red, and the head is marked with velvet-black bands.
According to the well-known German ichthyologist Dr. Hermann Meinken,
its behavior seems to be nearer that of an ordinary tetra than that of
Crenuchus spilurus.
Family Characidiidae
South American "Darters"
Most members of the Characidiidae, ali of small to extremely small size,
have evolved toward a life suited to the peculiar medium of river bottoms
and, sometimes, to the immersed roots of trees. The 'original members of
the family (of unknown line) have probably recent representatives in the
Characidiinae from southeastem Brazil, which still have a middle-water
life. The family, which attracted attention only rather recently, has been split
into subunits as follows:
111
-
Lebiasina astrigata. Photo by Or. Herbert R. Axelrod.
Lebiasina sp., possibly L. panamensis. Photo by Or. Herbert R. Axel-
rod of a speclmer'l from Rio Callma, Colombia.
Nannobrycon unifasciatus, form lacking caudal ocellus. Photo by Or.
Herbert R. Axelrod.
112
113
a. Dorsal fin short, usually with 10-13 rays (sometimes 14 in Jobertina rachovi);
small species mosdy adapted to life on river bottoms, the extreme reached
in the elongate body, fiat breast, padd1e-like pectoral fins, inferior mouth
and superior eyes of species of the torrents; lateral line complete or incomplete
... subfamily
CHARACIDIINAE
aa. Dorsal fin long, usually with 17or more rays; very small species, mosdy adapted
to life around submerged roots or falling leaveson the edge of small streams, with
a short body and the pectoral fins not gready enlarged; lateralline incomplete
or absent.
b. Teeth tricuspidate, with only one row on premaxilla; no maxillary teeth;
pterygoid teeth present; suborbital bones reduced; posterior fontanel present
. ... subfamily
BLACHOCHARACINAE
bb. Teeth conical, with two rows on premaxilla; maxillaryteeth present; pterygoid
teeth absent; suborbital bones absent; fontanels absent
. . . subfamily GBISLBRIINAE
-- ---- --
~. -
ti.
.
..
.-"'
..
"'
.
... ___AJtltiIIr
. . *~"'W..~ ~
,...
SUBFAMILY CHARACIDIINAE
South American Darters
This group has long been considered as comprised of only a few species,
with most of the specimens collected given the passe-partout name
Characidium fasciatum (the oldest named characidiid). Thanks to recent
studies, chiefiy those of the Brazilian expert Haroldo Travassos, the
complexity of the group has been revea1ed and the number of described and
still undescribed species may be estimated at 50 or more. Within the
Characidiinae, three genera may be recognized:
a. No maxillary teeth (pectoral fin with 3 to 5 unbranched rays and 6 to 12branched
on~s).
b. First dorsal ray dot rudimentary; lateral line incomplete; species from
southeastern South America, living in mid-water (with exceptions)
. . . Jobertina
bb. First dorsal ray rudimentary; lateral line complete; species not restricted to
southeastern South America, lsually living on stream bottoms
. . . Characidium
aa. 4-8 maxillary teeth (pectoral fin with 4 or 5 unbranched rays and only 4 to 8
branched ones) . . . K/ausewitzia
"
,.
-p---
:j.. "
----
Characidium species. Photo by Timmerman.
The trueJobertina livein mid-water and lookquite generalized. The other
ones look like tiny, regressed Characidium,and probably live on the bottom
in small rivers and brooks. Certain species have lost not only their lateral
line, but even their adipose fin, and they would probably not be classified
in the future within the genus Jobertina.
GENUS JOBERTINA
GENUS CHARACIDIUM
SOUTHAMERICAN DARTERS
Unlike most characoid groups, Characidiumand its allies have plenty of
characters on which to base a convenient and stable classification. These
characters, such as the pectoral and ventral rays, presence or absence of
scaleson the isthmus, position of the eyes and form of the teeth, correspond
with the way of life of the fish, depending on whether it lives on sand or
rocks, in fast running waters or slow moving ones, etc. However,
adaptations are rarely very restrictive in the teleostean fishes, and certain
species may tolerate quite a few different biotopes.
The genus Jobertina is still in a confused state, being not well defined.
It includes a well-delimited group of four or five southem species (J.
interrupta, J. lateralis, J. baht'ensis, J. rachovi, and J. theageri, respectively
from Rio de Janeiro, upper Paraguay, Babia, Paranagua, and Uruguay) and
another, composite Amazonian-Guianan group (J. eleotroides, from Guiana
plus a few other, still undescribed, species).
114
.\
Lcia Helena Rapp Pu - Daltlel
c, Po.,.l, 4:'a.iNPA .,
",000 - 11I - ""","zo.u _ 11,.."
115
.
,.
"
Nannostomus espei. Photo by Dr. Herbert R. Axelrod.
Nannostomus bifasciatus. Photo by Dr. Herbert R. Axelrod.
116
'''
~
- ""'\'':\'''' .
,. ' " .. . .: " ,., \
- "',..:". . < \, ~\\'.:~,:-. .'.,--
~f) " "", _ ,
1_, ~.__ , '. ,._.,~_:_
~
.
"..; .,
.',
. '. ,._<~'~~
~
,'~"{..., "', '.. ' .-~~'"
.',:."'" :', -~ . "-;:~.....
~" ': ~. . ":~::;:",
~,' :~ . . -~~
~, .. ~'" --~,
~, ~ - 'C''-'Y:'
Nannostomus espei spawning. Photo by H.J. Richter.
Male (above) and female (below) Nannostomus espei approaching
spawning site. Photo by H.J. Richter.
The complex genus CharacidiumwilI be treated here very briefly, by
means of the foIlowingartificial ky:
,...
a. First three pectoral rays unbranched, sometimes thickened.
b. First ventral ray unbranched.
c. Isthmus (pectoral region just behind the gill membrane) sca1ed
. .. C. fasciatum-group in the
broadest sense.
(This group inc1udesat
least 20 Guianan and
Amazonianspp., which are
mostly identifiedby the
structure of their teeth and
their proportions)
cc. Isthmus naked, or with 3 to 7 or more scales lacking (sometimes only one
scale lacking) . . . C.purpuratum-group inthe
broad sense. (This group
inc1udes 15 to 20 spp.,
mostly from northwest
and southeast South
America, very few from the
Amazon basin)
bb. First two ventral rays unbranched . . . undescribed speciesfrom
Rio de Janeiro
aa. First four pectoral rays unbranched and sometimes thickened.
d. First ventral ray unbranched . . . C. crandelli-group (2spp.
from Rio Branco and
Ecuador)
dd. First two ventral rays unbranched, at least on one side, and sometimes
thickened (isthmus usually naked) . . . C.b/ennioides-group (5very
specializedGuianan spp.,
for the most pan
undescribed)
.
4
Type of K/ausewitzla vintonl (after Eigenmann, 1912).
GENUS KLAUSEWITZIA
SOUTHAMERICAN DARTERS
This interesting genus, which differs from Characidiumin the presence
of maxillary teeth as weIl as in the smaIl number of branched pectoral rays,
has been recognized only recently. To the already known species K. vintoni
(synonymK.laterale) and K. ritae(the type species), one must add a curious
50 mm-long species which has been observed in nature by Dr. RoIf Geisler.
It is elongat~, green, and camouflages itseIf with the algae-like plants un-
dulating in the current. To add to this mimicry, it always stays near the top
of the narrow leaves, head upstream. Klausewitzia vintoni is quite similar
in meristic characters and it is not impossible that it has the same ecology.
SUBFAMILV ELACHOCHARACINAE
This curious group incIudes one genus with at least three extremely small
species living in the fine roots of plants growing in the banks of brooks.
They are shorter than most species of Characidium,and have a striped brown
pattern which may be homochromic with the surroundings.
Type of E/achocharax pu/cher. Original drawlng by Mr. Pablo Bravo In
1930,published for the flrst time by permission of Dr. G.S. Myers.
Possibly a new species of E/achocharax from Madeira basln. Photo by
Dr. Herbert R. Axelrod.
-
..
.
118
119
Nannostomus beckfordi. Photo by Harald Schultz.
Male Nannostomus beckfordi in full color. Photo by H.J. Richter.
120
Xanthochromic variety of Nannostomus beckfordi; usually it is called
"gold anomalus." Photo by Dr. Herbert R. Axelrod.
Another form of the Nannostomus beckfordi "complex." Photo by
Stanislav Frank.
121
Elachocharax pulcher, the first known species, lives in Venezuela.
E. georgiaelive~in small tributaries of the Rio Negro near Manaus,in the
Upper Amazon, as well as in the Rio Madeira basin, where a "giant"
specimen has been found (25 mm in standard length). A third species,
lacking an adipose fin, is still undescribed.
SUBFAMILV GEISLERIINAE
The single known species, Geisleriajunki, is even more aberrant than the
Elachocharaxspp. just mentioned. If not for the structure of the first and
last vertebrae, which are characteristic of a characoid, it could easily be
thought to be a killy. Its ecology is likely to be the same as that of
Elachocharax.
Holotype ot Geisleria junki, trom the Rio Madeira basin (atter Gery,
1973, in Vie et Milieu, with permission), and radiograph ot the speci-
men, made by Dr. W. Bom, Freiburg, Germany.
122
I
I
J
'.
Family Lebiasinidae
Penei I Fishes and Pyrrhulinins
The lebiasinids have many characteristics in common with the trahiras
(erythrinids). For example, almost nothing differentiates a Lebiasina
externally from an Erythn"nus. Both have the same body shape and head
form, both lack the adipose fin, and both are adapted to a life in poorly
oxygenated waters. However, certain features of their internal anatomy,
as well as different dentition (the lebiasinids usually have a short maxillary
bone, 2 rows of teeth, at least on the lower jaw and never the strong conical
or canine-like teeth of the trahiras), seem to be in contradiction: the
lebiasinids are presently considered as coming from a different stem than the
erythrinids.
Within the group, two subgroups are easily recognizable: (1) Lebiasininae
proper, and (2) Pyrrhulininae, which look like pigmy lebiasinids and inc1ude
the well-known pencilfishes and pyrrhulinins.
Technically, they can be separated as follows:
a. Median-sized fishes with a relatively long maxillary bone and tricuspidate teeth;
a patch ofteeth on palate; four branchiostegal rays (upper caudallobe often longer
than lower one; lateralline incomplete; adipose fin irregularly present)
. . . Subfamily LEBIASININAE
aa. Small fishes with a short maxillary, usually not functional, teeth conical (Pyrr-
hulina et aI.) or multicuspidate incisiform (Nannostomus et a1.), never tricus-
pidate; no teeth on palate; three branchiostegal rays
. . . Subfamily PYRRHULININAE
b. Mouth terminal; teeth multicuspidate (except the inner mandibular row,
which is composed of a few conical teeth); lateralline incomplete; caudallobes
equal or the lower one longer (in species which swim head up); adipose fin
irregularly present " . " Tribe Nannostomini
bb. Mouth upturned; teeth conical on both jaws; lteralline completely lacking;
upper caudallobe longer than lower one; adipose fin never present
. . . Tribe Pyrrhulinini
SUBFAMILV LEBIASININAE
This group c1assicallyinc1udes two genera, Lebiasina and Piabucina.
The characters which separate them, however, are variable and the genus
Piabucinais treated here as a synonym of the older genus Lebiasina.
.All species are from the northern part of South America, with the
exception of Lebiasinaintermedia, said to come from Santarem on the
Lower Amazon. Some, like L. bimaculata,have been introduced into other
countries in the fight against mosquitoes. The following key gives the main
points distinguishing the species:
123
Nannostomus trifasciatus. Photo by Dr. Herbert R. Axelrod.
Nannostomus marginatus. Photo by H.J. Richter. Above and below: A pair of Nannostomus marginatus spawning. Pho-
tos by H.J. Richter.
r
jl
CII
"
Lebiasina sp. from the Cauca Valley, 15 miles from Cali, Colombla.
Photo by Or. Herbert R. Axelrod.
j
~
a. 37scales in a longitudinal series and 9 in a transverse series; dorsal fin at mid-body
or dose to it (a vertical caudal spot; adipose fin usually present)
. . . L. erythrinoides, type of the
genus Piabucina
(Venezuela)
aa. 25-32 scales in a longitudinal series and 6-8 in a transverse series; dorsal fin
behind mid-body.
b. Depth of body 35.4.0 in standard length (usually 28 longitudinal scales;
branches of mandibles, as seen from below, forming a U; humeral and dorsal
spots present).
c. Length ofhead 4.2-4.5 in standard length; adipose fin present only in 5% of
the samples; caudal spot and median longitudinal band present
. . . L. bimaculata, type of the
genus Lebiasina (originally
from Peru and Ecuador)
cc. Length of head 3.7-4.25 in standard length; adipose fin usually present; no
caudal spot, a longitudinal band below median line
. . . L.lestae (Panama, N.W.
Colombia)
bb. Depth of body usually more than 4 in standard length.
d. A series of spots along mid-axis (depth of body 4.5-5.0and length of
head 4.0-4.75 in standard length).
e. Adipose fin present in only 10% of the samples; spots small and
numerous in from of dorsallevel; no dorsal spot
. . . L. multimaculata (N.W.
Colombia)
ee. Adipose fin usually present; spots large and roundish; usually a dorsal
spot . . . L. astrigata (Ecuador and
maybe Rio Meta in
Colombia)
.
J
126
I
I
~
dd. A longitudinal band (sometimes two) (adipose fin usually present).
f. Depth of body 4-4.6in standatd length; 10 or 11 anal rays.
g. 7 or 8 scalesbetween dorsal and ventral fins.
h. 27 longitudinal scales(branches of mandibles fortning a V)
. . . L. unitaeniata (Guyana)
hh. 28-32 longitudinal scales.
i. Length of head 3-4 in standatd length.
j. Longitudinal band often split up into irregular blotches
. . . L. panamensis(Panama)
j j. Longitudinal band continuous, though not well delimited
. . . L. boruca (Costa Rica)
i i. Length of head 4-4.33 in standatd length.
k. No dorsal spot; two lateral bands and a small caudal spot.
. . . L. aureoguttata(Ecuador)
kk. A dorsal spot; one lateral band and a roundish caudal spot
(branches of mandibles fortning a V)
. . . L. pleurotaenia (Venezuela)
gg. 6t scalesbetween dorsal and ventral fins (32 longitudinal scales;
depth of body 4.5 and length of head 4.25 in standard length;
lateralline going up to dorsallevel; humeral, dorsal and caudal
spots present, plus a lateral band) . . . L.intermedia (Santarem)
ff. Depth of body 4.7-5 in standatd length; 12 anal rays (branches of
mandiblesfortninga V; humeral and dorsal spotspresent,plus a
lateral band) . . . L. elongata (Upper Amazon
in the northem part of
Peru and south ofEcuador)*
SUBFAMILY PYRRHULlNINAE
Tribe Nannostomini
Penei I Fishes
The pencilfishes are among the best known characoids. Although
somewhat delicate in aquaria and difficult to induce to breed, they are very
popular among aquarists. Their fusiform shape, their brilliant colors and the
curious "head up" swimming of certain species amply justify their
popularity. All pencilfishes have avery small, always open mouth with the
anterior part of the mandibular bone movable over the posterior part,
which enables them to pick up very small prey on various substrata (mostly
plants, rocks and fallen branches).
According to a recent revision, only two genera are recognized within
the Nannostomini: Nannostomus and Poecilobrycon. t They are separable
on the basis of a character which is difficult to observe by the layman,
i.e. the presence or absence of a canal in the second circumorbital bone of
*The author lacks data concerning Lebiasinauruyensisfrom Venezuela.
t A still more recent revision, that of Weitzman and Cobb (March, 1975),
recognizes only one genus, Nannostomus.
127
--
'{
,
. ""
' ~~;.:'
~
- /. "
"''''',
~
..
""'~'-,
. " ,"
\ ""-"", / .
..
..
' ~ ~~"
'-,./ - .' l
/"
.;l'"
1
.I
. ':,
..~?
,
':'..,
,.
.'. ''.; .'
'~~~
''\
, '''~t
li
,"-
"
Nannobrycon unifasciatus. Photo by HaraldSchultz.
Nannostomus harrisoni. Photo by H.J. Rlchter.
Nannobrycon eques. Photo by HaraldSchultz.
Nannobrycon eques. Photoby H.J. Richter.
~
"'; .:.1.
,,' '."
i>~'j
. "'~.I
129
--.
the "cheek," which is associated with the length of the snout (distinctly
longer in Poecilobrycon).Actually, two species of the genus Poecilobrycon
(those belonging to the subgenus Nannobrycon)swim in a head-up position
and the lower lobe of their caudal fin is distinctly larger than the upper one.
These characters seem more convincing and the genus Poecilobrycon
(type-species P. harrisoni) is here abandoned in favor of Nannobrycon
(type-species N. eques),Poecilobryconbecoming a subgenus of Nannostomus.
f. Median stripe ending on caudal pedunc1e
. . . N. trifasciatus(upper
Amazon), the three-lined
pencilfish
ff. Median stripe extending posteriorly onto caudal fia
. . . N. erythrurus(Guyana;
middle and lower Amazon,
inc1udinglower Rio Negro),
also a three-lined pencilfish
ee. Adipose fia absent; 21-23 longitudinal scales
. . . N. marginatus(Lower
Amazon, Guyana and
Surinam, Colombia), the
dwarf pencilfish
GENUS NANNOSTOMUS
Subgenus Nannostomus
PENCILFISHES
In contrast to several other characoid groups, the seven or so species of
Nannostomushave distinctive diurnal color patterns which make them rather
easy to identify, at least if one does not bother with details. The term
diurnal is used on purpose, because all nannostomins have a so-called
nocturnal color pattern, which is usually composed of two or three blotches,
very different from the stripes of most species during the day.
The following key is based mostly on coloration:
a.5 oblique blotches on flanks, as if the fish were in a sort of permanent noctumal
pattem (adipose fia present) . . . N. espei (Guyana), the
barred pencilfish
aa. Diurnal color pattern composed of stripes along body.
b. Only one stripe somewhat below mid-axis of body (adipose fia absent)
. . . N. beckfordi (Guianas,
lower Rio Negro, lower and
middle Amazon basin)t
bb. Two or three stripes present.
c. 2' stripes, one similar to that of N. beckfordiand the other above mid-axis
ofbody. .
d. Adipose fia absent; a few scales with lateralline pores; anal rays iii, 9
. . . N. bifasciatus (Surinam and
French Guiana)
dd. Adipose fia present; no lateralline; anal rays iii, 8
. . . N. digrammus (Rio Madeira,
lower and middle Amazon,
and Guyana), the two-
striped pencilfish
cc. Three stripes, the two first like in the preceding species and the last one low
on abdomen.
e. Adipose fia usual1y present; about 26 or 27 longitudinal scales.
Despite the apparent simplicity of the above classification, the genus
Nannostomus is probably quite complex, actually actively evolving, some-
what in the manner of the Mrican cyprinodonts.* For example, there are
indications that the species exist by "pairs" which are almost in-
distinguishable externally. Two cases are reasonably demonstrated, the
N. digrammus-bifasciatuspair and the N. trifasciatus-erythruruspairo The
same situation is likely to occur in N. beckfordi and in N. marginatus, which .
are quite variable. Members of the pairs live in different localities
(allopatric forms), and their status as species can on1ybe ascertained on the
basis of their different chromosomes. This does not eliminate the
possibility of sibling (sympatric) species of each member of the pair, and
the variable chromosome formulae (karyotypes) give some indications of it
(J. J. Scheel, personal communication).t
Subgenus Poecilobrycon
The single species Nannostomusharrisoni, rarely seen in the aquarium,
looks much like an elongate Nannostomus beckfordi, with the following
coloration: a single stripe below mid-axis of body extending onto caudal
fin; anal and ventral fins blue with a red spot in the male, which also has
red spots on the caudal base above and below the extension of the stripe.
It swims horizontally and differs from a typical Nannostomus only in the
longer snout and the presence of the second infraorbital canal; it possesses
an adipose fino This species is found in Guyana and also in the Upper
Amazon and (probably) the Middle Amazon.
t A few forms, which may be sibling species or subspecies, can be recognized, for
example N. minimus and N. aripirangensis;N. anomaluscould be a synonym of the
nominate species; it is still a popular name of the Beckford's pencilfish, chiefly for
the xanthochromic strain, an artificial variety.
* For the evolution of Aphyosemion see J. Scheel, 1968. Rivu/ins of the Old World,
TFH Publications.
t In the revision cited in an earlier foomote, Weitzman and Cobb have just
described a new species, Nannostomusmarilynae, that looks somewhat like a small
N. trifasciatus with slightly different diurnal and nocturnal coloration.
130 131
Pyrrhulina of the P. brevis group. Photo by Dr. Herbert R. Axel-
rodo
Pyrrhulina sp., also of the P. brevis group. Photo by Dr. Herbert
R. Axelrod.
Pyrrhulina of the brevis group, the form generally called P.
IIlugubris. " Photo by Harald Schultz.
I
)
Pyrrhulina of the P. brevis group from Rio Aguaro. Photo by Dr. Her-
bert R. Axelrod.
Pyrrhulina laeta includes P. semifasciata as a synonym. Photo by
Harald Schultz.
133.
Poeci lobrycon
Nannobrycon eques,usualIy sympatric with the preceding species (same
localities) and resembling it, has only 24 or 25 longitudinal scales with not a
single one perforated, and the adipose fin is sometimes absent. The brown
longitudinal stripe, welI below the mid-axis of the body, also extends onto
the enlarged lower caudallobe; it is underlined by a series of spots Onthe
center of the scalesof the belIy and bordered above by a silvery band, above
which, on the upper part of body, there are two narrow lines. The two
oblique transverse bands of the nocturnal pattem of the nannostomins
are often superimposed on this longitudinal pattem, even during the day.
The anal fin is brown with a red spo~in the middle, and the brown lower
caudallobe is spotted with orange. This species is much less variable than
its cousins.
3
Tribe Pyrrhulinini
Pyrrhulinins
The pyrrhulinins are much lesswelIknown than the pencilfishes, and they
have never been revised (the author is informed that an important work was
in progress at the time this was written). They are surface fishes, living
mostly in smalI ponds together with the rivulins, but they are not able to
traveI on moist soil from pond to rivulet and so on, as do Rivulus species.
They have a relatively large mouth with acute teeth, as contrasted with the
pencilfishes, and some species (at least) are capable of ingesting ants that
have fallen into the water. They have no adipose fin, no lateralline and no
fontanels. Almost alIhave a plain colored body or are with a stripe and a dark
spot on the dorsal finoAt least some have a very elaborate reproductive
behavior, which is found mostly in more advanced groups, protecting their
eggs until hatching.
The 30 or so described species falI into three rather welI defined groups,
at least anatomicalIy, as folIows:
Dlstinctlons between circumocular bones In Nannostomus and Poe-
cilobrycon. Original drawlng by the author.
GENUS NANNOBRYCON
In this genus, whose species swim head up at a 45-75 angle, the lower
caudallobe is distinctly larger and is formed by 11 principal rays (instead of
9 in Nannostomusas welIas, actually, in almost alI characoids), and the snout
and second infraorbital are Poecilobrycon-like.
The two species have often been confused. Nannobryconunifasciatus,the
one-lined pencilfish, from the Middle and Upper Amazon basins, the Rio
Negro, the Upper Rio Orinoco and Guyana, has 28-30 longitudinal scales
with the mst 2-5 scales perforated; the adipose fin is apparently always
present. It has a longitudinal band similar to that of N. harrisoni and N.
beckfordi,which extends onto the enlarged lower caudarIobe in various ways
depending on the population. In specimens from Guyana the center of the
lobe is light (red in life) and there is an ocelIus above it, at the inferior part
of the upper lobe base (this population has been described as a distinct
species, N. ocellatus);in specimens from Brazil, among others, there is no
such ocelIus, or an indistinct one. As in the genus Nannostomus,a pair of
semi-cryptic species may be suspected.
a. Opercular membrane united to isthmus far forward; two complete rows of teeth
in upper jaw; maxillarybone straight or slightlycurved, never "S"-shaped, usually
toothed; nostrils dose to eachother . . . genus Pyrrhulina
aa. Opercular membrane united to isthmus far backward; one row of teeth on upper
jaw, plus usually one or two teeth behind it near the middle, not constituting a
r~al row of teeth; maxilla always toothed.
b. Body short, the body depth not more than 4 in standard length; fins without
filaments; outer edge of maxillary bone not "S"-shaped; nostrils dose to each
other . . . genus Copeina
bb. Body elongate, the body depth 4.25 to 6.0 in standard length; fins filamentous
in adults; outer edge of maxillary bone forming a double curve in the shape
of an "S", with a rudimentary apophysis; nostrils separated from each other
bya cutaneous bridge . . . genus Copel/a
134
135
Pyrrhulina vittata. Photo by Or. Herbert R. Axelrod.
Pyrrhulina spilota. A form with a short body approaching that of P.
vittata. Photo by Or. Herbert R. Axelrod.
L -
136
,.-- -
r--
-- ~
--Wp- . .
~A ..." ~'AD.~ -".
~'
.,." .
'~
Pyrrhulina sp. Photo by Klaus Paysan.
Pyrrhulina spilota. Photo by Or. Herbert R. Axelrod.
i
~
~
1
I
J
GENUS PYRRHULINA
PYRRHULlNAS
The charming dwarf fishes of the genus Pyrrhulina (the name means
"small bullfinch"), which are surface fishes with the mouth upturned, the
top of the head flat and without fontanels, and the upper caudal lobe en-
larged, are among the lesser known characoids. Not one species, except the
most recently named such as P. stoli and P. spilota, has been properly
described or even figured. It is hoped that the prospected revision referred
to above will put some order into the genus. Meanwhile, it is possible to
recognize several groups: (1) P. rachoviana-group; (2) P. brevis-group;
(3) P. laeta-group; (4) P. filamentosa-groupand (5) P. vittata-group. These
groups, perhaps quite artificial, are mostly based on external characters
such as color pattern (except for the first group).
(1) Pyrrhulina rachoviana:
Rachow's pyrrhulina from La Plata basin is the only Pyrrhulina with a
straight, un-toothed maxillary bone. In the other species, the outer
maxillary edge is always more or less curved and always bears some teeth
(usually 7 or 8 even in species considered as having a toothless maxilla, such
as Pyrrhulina vittata). It is a deep bodied species with a longitudinal dark
stripe from snout to caudal pedunde; its most colorful parts are the eye
and the base of the ventral and anal fins, which are red in life. This species
lays its eggs on a broad submerged leaf or, less commonly, on the bottom.
(2) Pyrrhulina brevis-group:
There is much confusion concerning this group, which is characterized
by a relatively deep body (body depth 3.4 to 4.25 in the standard length),
fewscalesin a longitudinal row (20-23) and a very short stripe which extends
onto the eye or usually to the operde.
In the author's opinion, there may be only one species, with slightly
divergent characters depending on the locality. In the Rio Negro around
Manaus the typical specieshas a short and well curved maxilla; in the south,
the subspecies P. brevisaustrale,from La Plata-Parana-Paraguay basin and
perhaps the Rio Guapore basin (with possible synonyms P. macrolepisand
P. stigmasemion),is slightly more elongate (body depth usually 4.0-4.25 in
the standard length versus 3.4-4.0 in P. brevisbrevis),with the stripe variable.
P. lugubrisis an intriguing form with the dorsal, ventral and anal fins black
edged and the maxilla more elongate and less curved than that of P. brevis
brevis. It may be a variety from the upper Amazon basin (and also Rio Meta
in Colombia), but it may also be a sibling species.
Pyrrhulina brevisis said to spawn on a flat stone.
(3) Pyrrhulina laeta-group:
P. laeta and associated forms are very dose to P. brevis. However, they
always have more longitudinal scales (23-28), the stripe extends onto the
138
Pyrrhulina species, from the Brazillan Amazon. Photo by Harald
Schultz.
second or third scaleor more and, especiallyimportant, the species have one
or two elongate spots on the predorsal region, easily seen from above.
Pyrrhulina eleanorae,from the Rio Ucayali, and Pyrrhulina stoli from
Surinam, Guyana and Upper Rio Meta (?), have the stripe extending up to
the second or third longitudinal scale, not more. They are extremely dose
to each other. However, a dose examination shows that P. eleanoraehas
slightly fewer predorsal scales (11) than P. sto/i (12-13). They may be two
geographic varieties of the same species.
Pyrrhulina laeta, from the Amazon basin, is a pain in the neck of an
ichthyologist. Its type, dating from 1871,lacks the anterior part ofthe body,
induding the head! Dne may believe tht, in such a case, it would be better
to destroy the specimen, and start off again with a new name and a new
description. The Rules of Zoological Nomendature forbid such an act: the
half-specimen may be useful in the future, if a new technique to charac-
terize the species, a chemical one for example, is discovered. We are pre-
sently left with the oldest valid name without a really good wayto attach it to
such-or-such a species! By elimination, the most "probable" species is the
one which has been named a fewyears later by Steindachner, P. semifasciata
(and, later on, P. maximaby Eigenmann). P. laetahas a stripe extending onto
the flank to the levei of the dorsal fino Its reproductive behavior is not
known.
139
,.,
Two photos showing a spawning sequence of Copeina guttata. Photos
by R. Zukal.
\
--,-
y
.."
~\
~
1 ,';" \ ~
.
.
..
:-
..
. .
1=-~
I
l~.
~~
lIA
..
.'
..
Copella of the nattereri-group (possibly C. eigenmanni). Photo
by Harald Schultz.
Copeina guttata. Photo by Dr. Herbert R. Axelrod.
(4) Pyrrhulinafilamentosa-group:
Pyrrhulinafilamentosa, the type species of the genus from the Guianas,
is a relatively large-sized species which resembles the genus Copellaby its
elongate shape, the body depth ofthe adult males attaining 4.25-4.75 in the
standard length (versus usualIy less than 4.25 in alI other pyrrhulinas except
beni). It has 26-28 longitudinal scalesand the usual stripe extends to the eye
(as in Pyrrhulina brevis).
Pyrrhulina beni, from Bolvia, may be its closest relative, also being
elongate (body depth 4.45 in the standard length), with 25 longitudinal
scales. The longitudinal band, like that of Pyrrhulina laeta, extends up to
the dorsallevel.
Nothing is known concerning the breeding of either species.
(5) Pyrrhulina vittata-group:
The two species forming this group are very characteristic, having a
spotted color pattern instead of a striped one as in the other species. They
are in a somewhatsimilarpositionas Nannostomus espeiis in comparison
with the other nannostomins.
Pyrrhulina vittata, the banded pyrrhulina, from the Amazon basin
including Rio Madeira, has a stripe extending to the opercle, 2 conspicuous
spots on the body and a transverse precaudal bandoIt is a smalI, deep bodied
species with few scales(20-22) and 3 rows ofteeth in the lower jaw(which is
unusual). Like P. rachoviana,it spawns on a broad submerged leaf.
Pyrrhulina spilota, of unknown origin, is an interesting novelty. It has
about the same pattern as P. vittata, but usualIy has one more spot and a
slightly more elongate body. It also has more scales (25 or 26). The method
of reproduction is not known.
GENUS COPE/NA
COPEINAS
In a certain sense, the one or two Copeinaspecies (if two species, they are
very closeto each other) are the least specializedofthe pyrrhulinins, and are
rarely found at the surface. They resemble lebiasinins the most with their
large size, their not very elongate body and their dorsal fin near the middle
of the body. Their maxilla is short, rounded, with none to five teeth.
The type species of the genus is Copeina guttata (synonym Copeina
argyrops),from the Amazon basin, which has 4 or 5 maxilIary teeth. In the
Peruvian Amazon it is replaced by a very similar form, Copeina osgoodi,
which can nevertheless be distinguished by the absence of maxillary teeth
(at least according to a cursory examination of the type specimen). Copeina
species hve a typical color pattern without a stripe but with a bright red
spot on most scales of the middle of the fianks, a red border on the ventral
and anal fins, as welIas on the lower caudallobe, alI standing out in contrast
to an overalI shiny blue tinto
142
t
.'
t
l
I
t
The mal e Copeina guttata fertilizes the eggs as they are being laid in a
depression in the sand. Photo by Ruda Zukal.
C. guttata (and probably also C. osgoodi, if the latter is a distinct species)
spawns in a hole in the sand, and the eggs are guarded by the male.
GENUS COPELLA
COPELLAS
The genus Copellawas erected by Dr. George Myers to include Copeina
species which have a double curved maxilIarybone forming a S (at least in
the adult male). At the time ofthe description of Copella(1956), oniy three
species were included in the genus. The examination of several species has
shown that alI forms previouslyknown as Copeina are indeed Copella
(except evidently the type species guttata and its close associate osgoodi).
ActualIy both sexes, when adult, have the maxilIa S-shaped, which is more
pronounced in the male, as welI as other characteristic features such as
nostrils separated by a dermal fiap, elongate body and filamentous fins.
(1) Copella arnoldi-group:
This group is characterized by a black stripe extending to the eye or
opercle. The species resemble Pyrrhulinafilamentosato such a point that the
aid of a binocular dissectingscopeis necessaryto classifya composite sample.
As far as known, Copella arnoldi, the welI known splashing tetra, is known
only from aquarium specimens, said to come from the lower Amazon. It is
moderately elongate (body depth about 4.25-4.75 in the standard length)
!
I
143
\$.. . ~
~ ~.
f'J&
.(:f"
I
t
\,
,/
""~
Unidentified Pyrrhulina. Photo by Dr. Herbert R. Axelrod.
Unidentified Pyrrhulina. Photo by Dr. Herbert R. Axelrod.
Copella nattereri. Photo by Dr. Herbert R. Axelrod.
----
.
144
.....
!li.. ..~ ...
1 :--- ~ "',
V
- ~..v"
~if,
.,.
~
<i.V~>~
~.
.
-.
.A':'I:
\
II
-"'I.
"\
'"
~.
,
,
,
..'l
",:, -.(;
"of'f<
<
.~
......
II"""'P-.) ,
..~<.., .. .
' .. '.'".....-
11". " IL
,.
~"'
\... " :v',\
, "'~"~~
\
~: .,~
~
>fj
'-
".. \
"
...
'tj
Copella nattereri. Photo by Dr. Herbert R. Axelrdd.
145
.
.,
I
.
i
"
..
'*
t fi_
(
and very graceful with its red, elongate fins and shiny scales. The pair
spawns on a broad leaf above the surface. The male remains under the leaf
and moistens the eggs and embryos with "splashes" of its tail.
Copellacallolepisis considered as a synonym of C. amoldi or C. nattereri
by some authors.
Copellacarsevennensis, from the Guianas, can be distinguished from C.
amoldi as follows: it has more scales (26 versus 23 or 24) and the dorsal fin
is slightly nearer the head than caudal fin (not so in C. amoldi).
(2) Copellanattereri-group:
This group is characterized by a more or less distinct longitudinal band
to the caudal fin and a slightly more elongate body than the preceding
copellas.
Comparison between the heads of a Pyrrhulina (P. filamentosa, left)
and a Copella (C. carsevennensis, rlght). Original drawing by the
author.
146
1
Male paratype of Copella compta. Original drawing by Mr. Pablo Bravo
In 1930, published for the flrst time with permlssion of Dr. G.S. Myers.
Copellanattereri,from the Rio Negro and the Middle Amazon, has a pale
spot on each scale, a dusky band along the body and a black triangular spot
at the base of the lower caudal lobe, which seems characteristic of this
species.
Copellametae, from Rio Meta, and Copellaeigenmanni,from Guyana and
Para, are to be distinguished from the preceding by the more distinct
longitudinal band and the somewhat more numerous scales (23 or 24 in a
longitudinal series and 14 in predorsal series, ver~us 20-22 and 12, respec-
tively). The two forms (distinct species ?) are extremely dose to each other
and differ in the dorsal fin position, which is more posterior in C. metae.
Pyrrhulina nigrofasciatais probably synonymous with the latter species.
Nothing is known about the reproductive behavior of these species, which
are quite difficult to identify in the aquarium.
(3) Copellacompta-group:
The two species of this group, gain extremely dose to each other,
"represent the acme of gracefulness reached by characins of the Pyrrhuli-
ninae" (G. S. Myers). The body depth attains 5.5 times in the standard
length, and the fins in the adult males are very elongate.
Copellacompta, from the upper Rio Negro, has a color pattern similar to
that of C. nattereri,at least judging from preserved specimens. Its coloration
in life is not known. The rainbow copella, Copellavilmae, from the upper
Amazon not very far from Tabatinga, is of a magnificent wine-red color with
some metallic blue spots on the flanks. Technically it differs from C. compta
in having fewer longitudinal scales (24 or 25 versus 26) and fewer predorsal
scales (14-16 versus 17), about in the same manner as C. nattereri differs
from C. metae, proportionately speaking.
The reproductive behavior of neither species is known.
147
,
~
148
1..
.
J
,
Copel/a of the arnoldi-group. Photo by Hilmar Hansen, Aquarium Ber-
lin.
Another male Copel/a of the arnoldi-group. Photo by Or. Herbert R.
Axelrod.
. --.-
-
I
,/'
",,/
,...
j
--
I -;11'"
~
\
Copel/a sp. from Rio Negro, possibly Copel/a compta. Photo by Or.
Herbert R. Axelrod.
Copel/a vilmae, the rainbow Copel/a. Wild mal e in fuI! color. Photo by
Harald Schultz.
149
Family Anostomidae
Headstanders and Leporins
The Anostomidae, induding among others the welI-known leporins and
certain headstanders (not the chilodins), are a very satisfying group
from the taxonomists' point of view: they are well dassified, with a reason-
able number of fairlywelIknown genera, and the limits of the higher groups
are rather sharply defined. Only at the species leveI do the difficultiesbegin:
no one knows, for example, how many species of LeporinusrealIy do existo
The principal characters of the Anostomidae are:
-body usually elongate (except in Abramites)
-anal fin short, with less than 10 branched rays (except in Abramites)
-gill membrane united with the isthmus
-anterior nostril forming a tube
-teeth in one series of only 3 or 4 on each jaw, usually incisiform, firmly set
into the bone; no teeth on maxilIa, none on palate.
Two groups are recognizable within the AI,1ostomidae,the subfamily
LeporelIinae with the single genus Leporelluswhich has the nostrils dose to
each other and the caudal fin scaly, and the subfamily Anostominae, with
several genera which usualIy have the nostrils set apart from each other
and the caudal fin naked.
SUBFAMIL Y LEPORELLlNAE
GENUS LEPORELLUS
Apart from the characters given above, the species of the genus
Leporellusare like those of Leporinus,with a dentition somewhat similar to
that of subgenus Hypomasticusof the latter genus. Seven forms have been
described: L. vittatus, L. pictus, L. maculifrons, L. timbore, L. cartledgei,
L. retropinnis, and L. sexdentatus. L. maculifrons, L. timbore, and L. cartledgei
are probably synonyms of L. pictus. They are very dose to each other and,
apart from Lepore/lus retropinnis from the Rio Piracicabia in southeast
Brazil (which has the dorsal fin at midbody instead of welI in advance of
midbody), they may belong to the same species. They are recognizable on
first sight owingto the peculiar oblique bands oftheir unpaired fins, strongly
reminiscent of those of certain Prochilodus (insignis-group). Lepore/lus,
although everywhere rare, inhabits alI the principal South American basins,
i.e. Orinoco, Amazon and Parana-Paraguay.
=-
SUBFAMILY ANOSTOMINAE
Leporins and Headstanders
The genera of this important group are welI defined by a few anatomical
characters, chiefty of their mouth and dentition, as folIows:
a. Mouth neverupturned.
b. Teeth incisiformwith2 to 4 or 5 cusps(bodyelongate)
. . . Schizodon
bb. Teeth withoutcuspsor simplynotched.
c. Bodynot veryelongate;teethusuallyrather thick,normallytruncate.
d. Bodynot veryhigh, not compressed;analfinwithlessthan 10branched
rays; no post-ventralkeel; most speciesswimin normal (horizontal)
position . . . Leporinus
dd. Body high and compressed; anal fin with more than 10 branched rays; a
post-ventral keel; species swim in head-down position
... Abramites
cc. Body very elongate, cylindriform; teeth thin, with irregular, finely denti-
culated, cutting edges (convex above and concave below)
. . . Rhytiodus
Head and teeth 0# a generalized anastomld, Leporlnus fasclatus, type
0# the genus Leporlnus. Original drawlng by the author.
aa. Mouth more or less upturned.
e.3 or 4 teeth on each side of lower jaw.
f. Body rather deep; upper teeth sometimes tricuspid, lower teeth
rounded, as in some Leporinus (mouth upturned, but the opening
not quite vertical) . . . Anostomoides
ff. Body elongate.
g. Teeth tri-to multicuspid, at least in upper jaw (Anostomus).
h. Mouth with opening not quite vertical, as in Anostomoides;
rather large size; upper teeth with 3 or 5 cusps, lower teeth
truncate when adult . . . Anostomus (Laemolyta)
150
151
I
~
Schizodon fasciatum adult collected at Solimoes, between Coari and
Jutica. Photo by Or. Herbert R. Axelrod.
Leporellus vittatus. Photo by Aaron Norman.
Or. Martin Brittan preserving a collection ot tishes, mostly anosto-
mids, trom Tete, Brazil. Photo by Or. Herbert R. Axelrod.
152 153
hh. Mouth strictly superior; small size (except a few forms); teeth
usually bi-or tricuspid on both jaws
.' . . Anostomus (Anostomus)
gg. Teeth usually rounded at the tip or at most bifid.
i. 4 teeth on each side of both jaws, usua1lyunicuspid, the
lower symphyseal pair not greatly enlarged
. . . Synaptolaemus
ii. 4 teeth on the upper jaw, bifid; only 3 teeth on lower jaw, the
symphyseal pair very long, protruding vertica1lyout of mouth
. . . Sartor
ee. 0nly one tooth on each side of lower jaw near symphysis, very long,
spoon-shaped at its end and extending above the snout as in Sartorj
lower jaw extremely narrow and slender
. . . Gnathodolus
A third group is best characterized by the slightly inferior mouth together
with the colorationas in S. isognathum: it is composedof onlytwoverysimilar
species, Schizodon nasutum from Matto Grosso (and maybe from Rio Sao
Francisco) and Schizodon rostratumfrom Rio Paranahyba.
Finally, the systematic postion of Nanognathus borel/ii, which has a
dentition similar to that of a Schizodon, is not dear. Some authors have
identified specimens of a Schizodon from Paraguay as Schizodon borel/ii.
However, in the original description of N. borel/iiit s expressly stated that
the nostrils are dose to each other and the gill membrane is free from the
isthmus, which seems to exdude Nanognathusfrom the Anostomidae. On
the other hand, LeporinusplatycephalusMeinken may be a Schizodon. AlI
Schizodon species attain a rather large size (maximum 260 mm in standard
length). In certain places such as Manaus, where they are rather abundant,
they are found in the fish market together with members of the next genus.
GENUS SCHIZODON
The species of this genus do not differ extemally from most Leporinus.
However, once the terminal mouth is examined, the pluricuspidate teeth,
forming a continuous crenulate cutting border, are quite characteristic. It is
much more difficult to distinguish the species, and the genus is certainly
in need of a revision.
The best known species, and the one whichis commonlyseen, is Schizodon
fasciatum from the Amazon basin; it has the median teeth with 4 cusps,
about 8+1+13 gill-rakers and 4 vertical black bands on the body, plus a
caudal spot which is well separated from the last band located above the
ana1fino
In Schizodon vittatum, also from the Amazon basin (Rio Araguaia), the
last transverse band is joined with the caudal spot and the center of each
scale is paler than its border, giving the impression of longitudinal dark
stripes between scale rows. It is likely to be only a color variation of S.
fasciatum. Another form, which also may belong to S. fasciatum sensulato,
is Schizodon dissimilefrom the basin of the Rio Paranahyba. It is best
identified by its elongate body and lack of a caudal spot.
In a second group of species we may indude Schizodon corti, S.
isognathum, S. kneri, and S. platae, which differ from the S. fasciatum-group
in having 5 or 6 transverse scales above and below the lateralline instead of
4. Schizodon corti, described as a subspecies of S.fasciatum, seems indeed to
be the representative of it in Venezuela, having exactly the same color
pattem; however, apart from the transverse scales, it differs in tooth struc-
ture, never having as many as 4 cusps on either jaw, but only 2 or 3.
Schizodon isognathum from the Rio Paraguay Basin, S. kneri from the Rio
Sao Francisco basin, andS. platae from Rio de Ia Plata (the latter two species
perhaps synonymous with the former) have a different coloration, having
lost the lateral marks but retaining the caudal spot. These forms are not
very elongate (depth about 3.66 times in the standard length).
GENUS LEPORINUS
LEPORINS
This is one of the largest genera of characoid fishes (together with
Curimata, Prohilodus, Hyphessobrycon, Hemigrammus, Astyanax and
Moenkhausia), and one ofthe least known. In typical Leporinus(as pointed
out by G. S. Myers) "the teeth are heavy and thick at the base, those of
upper jaw arranged in a step or stair-like pattem, the median pair set far
forward and largest; each tooth which followsis set as much behind as to the
side of the preceding tooth . . .and the cutting edge is approximately trans-
verse to the long axis of the fish; the lower teeth are slanted forward to the
median pair which is largest . . . with an oblique, usually concave cutting
edge."
The genus can be split into four subgroups as follows:
a. Mouth terminal or subinferior.
b. Lower median pair of teeth not developed into laterally compressed tUsks.
c. Small size, body rather compressed with ventral region keeled; mouth
terminal; 3 upper teeth only, bifid, not arranged in a step-like pattern (or
very feebly so)j gill membrane united far forward of isthmus, thus forrning
an acute angle . . . subgenus Leporinops
(2 species ?)
cc. Usually rather large size; body not very compressed, with ventral region
rounded or flattenedj mouth terminal or subinferior; 3 or 4 upper teeth
arrangedin a step-likepattern(seeabovedefinition) j gill membrane united
back to isthmus, thus forming an obtUseangle
. . . subgenus typical Leporinus
(several species)
bb. Lower median pair of teeth developed in adults into laterally compressed tUsks
(rather large size; ventral region somewhat flattenedj 3 upper incisiformteeth,
154
155
""
Schizodon fasciatum juvenile. Photo by Harald Schultz.
Leporinus of the pellegrini-alternus-group (Upper Amazon form).
Photo by Dr. Herbert R. Axelrod.
156 .
" .
""
~.~
"
Leporinus of the pellegrini-alternus-group. Photo by Klaus Paysan.
Young Leporinus of the fasciatus-group. Photo by Dr. Herbert R. Axel-
rodo
+~'"
"'*"
1-
157
not arranged in a step-like panern; gill membrane united far back to isthmus)
. . . Subgenus Myocharax
(one species)
aa. Mouth inferior, visible on1y from below, the snout almost vertical (small to
medium size; body and gill membrane like that of Leporinustypica1; 3 or 4 upper
teeth vertica1lyset, not arranged in a step-like panern)
. . . Subgenus Hypomasticus
(at least 5 described
species)
SubgenusLeporinops
This very small group of leporins with a slight ventral keel has Leporinus
moralesi as the type species (a name given by Fowler to replace Leporinus
muelleri, which was preoccupied). It is a small fish (less than 100 mm in
standard length), an inhabitant of the upper Amazon basin, with a color
pattern formed by three elongate spots ending in a longitudinal band extend-
ing to the top of the caudal rays, resembling that of Leporinus agassizi, etc.
Another species probably belonging to the same group is Leporinus jamesi
from the middle Amazon basin. It has no longitudinal pattern but only a
series of bars across the dorsum like in most Leporinus (chieBy when young).
Subgenus typical Leporinus
Leporins
It is beyond the scope of the present book to describe all the known
species (50 or 60) of this large group and to give their taxonomic characters
(which ar not even precisely known for each species). For want of a
c1assificationbased on well:'studied features, we may artificially split the
genus Leporinus into a few types of color patterns: (1) species with a
transverse pattern; (2) spotted species; (3) species with a longitudinal
pattern; and (4) plain species without a conspicuous pattern.
(1) Species with transverse bands on body:
This group, which inc1udes the type species of the genus, Leporinus
fasciatus, may in turn be divided into subunits:
(a) Speci~ with 4 bands (3 bands on body, plus a precaudal band):
This group inc1udes Leporinus trifasciatus, common in the middle
Amazon basin (sold on the Manaus market), and Leporinuswolfei, a c10se
form from the Upper Amazon, which is deeper bodied (depth 3.1-3.33
versus 3.3-3.7 in the standard length) than L. trifasciatus with fewer
longitudinal scales (40 versus 42 or 43).
(b) Species with 7 or 8 transverse bands:
This group inc1udes Leporinus conirostris from Rio Parahyba,
Leporinusoctofasciatusfrom Santa Catarina State in Brazil and Leporinus
latofasciatusfrom Rio Orinoco with 8 bands. These 3 species are poorly
Leporinus jamesi (above) and Leporinus mora/esi (below) (both after
Borodin, 1929).
known. Leporinus pellegrini, L. alternus and L. paralternus, from the
Guianas and the upper Amazon, have 4 broad bands alternating with
4 narrower ones. Leporinspearsoniand L. yophorus, from Colombia and
Bolivia respectively, strongly resemble each other, having 8 transverse
bars, but 2 or 3 ofthese are united (at the lateralline leveI)to form a Y.
(c) Species with 9 or 10 transverse bars in the adult (fasciatus-group):
Several names, L. fasciatus, L. multifasciatus, L. fasciatus altipinnis,
L. fasciatus tigrinus and L. Jwlostictus,probably apply to the same rather
variable form, whose transverse bars are quite numerous when young,
but some uniting during growth (about fourteen becoming ten). Leporinus
aifinis, from the state of Para, is perhaps different, having only~9bars and
the anal fin rounded (instead of pointed). The forms of the L. fasciatus-
group are quite abundant in the Guianas (type localityof Leporinusfascatus)
as well as in the Amazon basin.
(2) Species with spots on the body:
This group, which inc1udes no less than 30 described species, is
probably the least well known of alI leporin groups. Like the preceding
group, it can be split quite unnaturally into subunits depending on the
number of spots, which are usually large, round, and situated astride on the
lateralline. Smaller supplementary spots or dorsallines may be present in
158 159
r
.
Leporinus fasciatus. Photo by Harald Schultz.
Leporinus ot the fasciatus-group. Photo by Harald Schultz.
160
1
Adult Leporinus fasciatus tromTete. Photo by Dr. Herbert R. Axelrod.
Unidentitled anostomid trom Venezuela, posslbly ot the genus Lae-
mo/yta. Photo by Dr. Herbert R. Axelrod.
I
\L
161
,.
some species, at least when young, this being a usual pattern in the
spotted Leporinus. It cannot be taken into consideration except for the
so-calIed "maculatus-group" which seems to have retained at the adult
stage its juvenile coloration. Certain forms may be distinguished thanks to
their very elongatebody (this is associatedwith horizontalIyoval spots).
(a) Species with one or two spots on body when adult (no caudal spot):
There is one species with oniy one large spot at the dorsal fin leveI, i.e.
Leporinuslacustrisfrom Sao Paulo State in Brazil. It is one of the deepest,
largest scaledspecies. There is also a pair of Amazonian species (Leporinus
bimaculatus,L. nattereri) and one from Ceara (L. piau), alI with 2 spots,
which are in need of revision. It is possible that the 3 names cover a
single entry. On the other hand, the dose nostrils described in L. nattereri
would characterize a very distinct species, even a distinct genus, whereas
Leporinuspiau may eventually be found to be the young of L. reinhardti.
(b) Species with three or four spots on body when adult (induding a
caudal spot):
The dassification of the L. friderici-group is one of the worst Chinese
puzzles of characoid systematics. For brevity, this group will be treated
here by means of a tentative key to their identification. It is more than
probable that this key will soon be outdated.
Unidentified Leporinus of the friderici-group. Photo by Harald Schultz.
...
162
T
{
.
~
C
I
I
,
(
4
Key to the jriderici-group of Leporinusspp.
a. Body depth usually less than 3.66 times in standard length (friderici-group
sensustricto).
b. Less than 40 lateralline scales.
c. 4/3-1- or 4 transverse scales.
d. Depth about 3 in standard length.
e. 35 lateralline scales; head 3.5 in standard length; anal fin with 10 rays,
rounded, reaching caudal fin when depressed
. . . L. bahiensis(Bahia)
ee. 37 lateral line scales; head 4.2 in standard length; anal fin with 12 rays,
concave, not reaching caudal fin when depressed
, . . . L.steindachneri (Rio
J equitinhonha)
dd. Depth 3.4-3.66 in standard length (36-39 lateralline scales; head 4-4.2 in
standard length; anal with 10-11 rays, concave, not reaching caudal fin
when depressed; teeth pointed, 3-4/4) ... L. copelandi (RioParahyba)
cc. 5-6/5-6 transverse scales.
e:'4lateralspots . . . L. leschenaulti, synonymL.
parae(Oyapockand Para)
ee~3 lateral spots (inc1udingcaudal spot).
f. Head 3.33-3.4in standard length (depth 2.85-3.15 in standard length;
35-38 lateralline scales; 12 predorsal scales; teeth slightly notched,
4/3 ?) . . . L.subniger (Colombia)
ff. Head 3.8 in standard length.
g. Teeth slightly notched, 3/3; 35 lateralline scales; 10-11 predorsal
scales; 14 peduncular scales . . . L. reinhardti (RioS.
Francisco)
gg. Teeth truncate, 4/4; 36-38 lateralline scales; 10 predorsal scales;
16 peduncular scales . . . L.friderici,syn.L. acutidens
(Guianas, Amazon, probably
other basins; possible
subspecies: L. spilopleura,
L. ecuadoriensis)
bb. 40-43 lateralline scales.
h. 5/5transverse scales; body depth 3.33-3.5 in standard length
(sometimes more, or sibling species ?)
. . . L. muyscorum (Rio
Magdalena and Rio Atrato
in Colombia)
hh. 6-1--7-1-/5t-6transverse scales.
i. Body depth 2.9 in standard length; 6/5-1- transverse scales;
3/3 teeth . . . L.obtusidens, synonyms
L. pachyurus, L. brachyurus
and ? L. aguapeiensis
(La Plata and S.E. Brazil
to S. Paulo)
ii. Bodydepth 3.5-3.66in standard length; 6-1--7-1-/6transverse
scales; 4/4 teeth ? . . . L.silvestrii (Coxipo,
Mato Grosso)
163
,...----
Leporinus
steyermarki. Photo by
Harald Sehultz.
164
.....
Leporinus sp.
(nattereri-group?).
Photo by
Harald Sehultz.
---
~:,~?
'"
.~,~ ~
~
Leporinus of the
so-ealled "maculatus-
group" (Leporinus
granti aueL). Photo
by Harald Sehultz. jl
li
J
t
Leporinus agassizi. Photo by Harald Sehultz.
Leporinus melanopleura? Photo by Dr. Herbert R. Axelrod.
165
aa. Body depth more than 3.66 in standard length (cylindriformis-group).
j. 37-41 lateralline scales (see also couplet h., L. muyscorum).
k. Body depth 3.7-4 in standard length
. . . L. klausewitzi (Amazon
basin)
k k.Bodydepth4.5-4.66in standardlength
. . . L. niceforoi (Colombia)
j j. 44 lateralline scales (body depth 4.75 in standard length)
. . . L. cylindnformis(Lower
Amazon)
(c) Species with severa! spots on body at alI ages (so-calIed maculatus-
group):
The description of this group is hampered by a delicate problem of
nomenclature: the species called L. maculatus by ichthyologists and
aquarists (see for example page F.377.00 in the Looseleaf Edition of Exotic
TropicalFishesby Axelrod et aI.) is certainly not the one described by Mller
and Troschel under this name (it is not the proper place here for a nomen-
clatorial change; it sufficesto say that L. maculatusis in reality what is calIed
Color pattern of three forms of spotted Leporinus. Top to bottom:
L. (Hypomasticus) "maculatus" auct. (note inferior mouth), "granti"
auct., and steyermarki (schematic).
166
I
D
11
t
II
I
~
/
I
f
r
at present L. pellegrini). The situation is even worse than one may believe
because L. maculatus was applied to at least two different species belonging
to 2 subgenera; one with an inferior mouth and 3/3 teeth (see description
of the subgenus Hypomasticus below), and one with a terminal or scarcely
sub-inferior mouth. For the last one, which occurs in the Guianas and
the Amazon, fOUInames are available: L. megalepis (at least part ofthe types),
L. marcgravii, L. grami and L. badueli. A re-examination of alI megalepis-
material is necessary to determine which name is valido
A final species with several spots on its body, L. steyermarki from
Venezuela, differs from the preceding (which will be called here
provisionally Leporinus grami) as noted in the following key:
a. 3oval or almost rectangular spots, first under dorsal fin, second midway to caudal
fin, preceded by two smaller, less well delimited spots, and the third, smallest, at
end of peduncle; a series of 3 smaller spots above main series, beginning well in
front of first large spot; an oblique series of small spots belowmain series, usually
moreor lessobscure;a series of blotches or bars on dorsum; usually 16 circum-
peduncular scales . . . Leporinus "granti"
aa. 3 spots of main series round; spots of series above confused with bars across
dorsum; spots of series below only 3, the last one very small (color pattern
seemingly intermediate betweenfriderici-type and "maculatus"-type); 12 circum-
peduncular scales . . . Leporinus steyermarki
(3) Group of species with a longitudinal pattern:
The seven or so species with one or more longitudinal dark bands are
slightly better known than the preceding ones. They can be distinguished
as follows:
a. Only one longitudinal band along axis of body (see also Leporinus moralesi,of
subgenus Leporinops).
b. Longitudinal band beginning at dorsal fin leveI.
c. Depth 3.2-3.45 and head 3.7-4 in standard length; 38-40 lateralline scales,
5/5 transverse scales . . . L.agassizi (synonym
L. semivittatus)(Amazon
basin)
cc. Depth 4.5-4.55 and head 4-5 in standard length; 40-42 lateralline scales,
5 or 6/5 transverse scales . . . L. nigrotaeniatus(synonym
L. margaritaceus) (most1y
Guianas; occurence in the
Amazon not proved)
bb. Longitudinal band complete, beginning behind opercle (depth 3.5-4.0 in stan-
dard length; 35-38 lateralline scales; teeth 3/4).
d. 4/4 tranverse scales . . . L. melanopleura (Bahia)
(squamation said to be
3!/35/2 in L. macrolepidotus,
an uncertain species from
Rio de Janeiro)
dd. 5/5 transverse scales . . . L. taeniatus (RiodasVelhas)
167
~-
.~
u
Leporinus friderici. Photo by Or. Herbert R. Axelrod.
Unidentified Leporinus, probably close to L. jamesi. Photo by Or. Her-
bert R. Axelrod.
-- -
. ,~
-~{,
~
~
.~
, ,
~"'J
~-"-~
168
--,r-
,
/'
8
Leporinus sp., probably of the nattereri-group. Photo by Or. Herbert R.
Axelrod.
Leporinus "maculatus" auct. Photo by Or. Herbert R. Axelrod.
169
aa. Three longitUdinal bands on flanks plus one dorsal bando
e. Medium sized fish, with largest size recorded 175 mm standard length;
depth 3.5-4.0 in standard length; teeth 3/4; profile of head straight over
eyes; dorsal fin midway between snout and adipose fin; no black spot at
pectoral fin base . . . L. striatus(Paraguay,
Uruguay, Mato Grosso,
Sao Paulo, ? Bolivia,
Colombia and Ecuador)
ee. Large sized fish, with maximum size recorded 400 mm standard length;
depth a little more than 3 in standard length; teeth 4/4; profile of head
slightly convex over eyes; dorsal fin midway between snout and a point
between adipose fin and beginning of caudal fin; a black spot at
pectoral fin base . . . L. areus (Guianas,
Venezuela, Upper Amazon
and ? Para)
.
(4) Group of species without conspicuous color pattem when adult, or
with only a caudal spot (certain species being known only from old
collections in museums may have some coloration when alive):
Unidentified species of Leporinus. Photo by Timmerman.
a. Mouth strictly terminal.
b. Body depth 3.9 and head-length 4.55 in standard length; 6 or 6!-/42/5-5!-
scales . . . L.elongatus (Buenos Aires)
bb. Body depth 3 and head-Iength 3.5-4.0 in standard length; 4!- or 5/35/3!-or 4
scales; a peduncular spot . . . L. melanostietus (Oyapokin
French Guiana)
aa. Mouth subinferior (species being at the limit of the subgenus Hypomastieus).
c. Body depth 4.15 in standard length; 4/4 teeth (5,38-41/5 scales; lips thin;
scales edged with dark) . . . L. brunneus (RioNegro)
CC.Body depth 3.0-3.75in standard length; 3/3 teeth (4!/37/3! scaks; lips thick;
no caudal spot when adult) . . . L.erassilabris (Rio
Jequitinhonha)
~
I
Subgenus Myocharax
The single species, Leporinus desmotes from the Rupununi River in
Guyana, has an elongate body, a long tubular anterior nostril, a.subinferior
mouth with thick lips, and 3/3 peculiar teeth: the upper ones directed
forward, somewhat obliquely, the lower ones also directed forward, almost
horizontally, the median. pair forming long tusks abutting against the
superior ones. In contrast with the lower teeth of Leporinusfasciatus, a
species which also develops "tusks" when old, these teeth are laterally
compressed. The color pattem is strikingly similar to that of Abramites
and Synaptolaemus.
Subgenus Hypomasticus
The speciesofthis peculiar group have a strictly inferior mouth, the upper
teeth clearly vertical or even pointing backward. Certain species eat fresh-
water sponges, but so do other leporins with a terminal mouth ("L.
pellegrini" auct. for example), as well as anostomids with a superior mouth
(Sartor respectus,for example). Although only 4 or 5 species have been
described, it is the author's feeling that a number of forms are to be dis-
covered in the future (at least 5 new species are already in collections).
Curiously, representatives ofthe subgenus Hypomasticusare usually mimics
of a form of the nominal subgenus Leporinus:for example L. (H.) despaxi
copying L. arcus,"L. (H.) maculatus"(auct.) copying L. granti, and so on.
As mentioned above, Leporinus"maculatus" (in the sense ofmost authors)
presently has no officialname. It is found in the Guianas as well as in the
southem Amazon basin and can be distinguished from its "model,"
Leporinusgranti (also a tentative name), nct only by its small size, inferior
mouth and some meristic characters, but aiso in details of color pattem,
notably the lower row of small spots which are well separated and not fused
171
,
Newly laid eggs (below) and developing embryo (above) of Leporinus
"maculatus." Photo by H. Azuma.
Leporinus "maculatus." Above: one moll;th old juveniles of 15-16 mm.
Below: four months old and 30 mm in length. Photos by H. Azuma.
~
.--
\
"
7h
.
'"'I :\.~
~.
.."",-
A\\ ,~. /
'H'\<y. .. . C
~
~
/..'. ,---
\'~~. ~--
.....
j
as in L. grami. L. "maculatus," one of the most beautiful of the leporins, is
unfortunately quite rare. Otherwise it would have become very popular in
the aquarium hobby. It has rather recently reproduced in ao aquarium,
however, and the present situation may change rapidly.
Aoother striking, but very rare, Hypomasticus, L. despaxi, also a small
form, cannot be told apart from young L. arcus or L. stn'atus by color pattern
alone. It is restricted, as far as we know, to Guianan rivers: the Maroni,
Mana, and Oyapok. .
The other already described Hypomasucus are plain or with a faint caudal
spot. They are aiso of smaIl size (135 mm maximum standard length),
extremely rare,and restrictedto riversof southeastern Brazil.They may be
identifiedas follows:
a. 36or37lateralline sca1es; depth 4 in standard length; 4/4 teeth; no spot on bocJt
. . . L. mormyrops, type ofthe
subgenus, synonym ,
L. thayeri ? (Rio Parahyba,
etc.)
aa. 34lateralline sca1es; depth 4.75 in standard length; 3/4 teeth; a faint spot at end
of caudal peduncle . . . L.garmani (Rio
Jequitinhonha)
(see also two dose species
with subinferior mouth
from the same region:
L. conirostris and
L. crassilabris).
174
:I
..
..
I
GENUS ABRAMITES
HEADSTANDERS
This genus has the same characteristics as the genus Leporinus (chiefty
its subgenus Leporinops) with the exception of a higher body, a post-ventral
keel and a longer anal fio. The most common forms have the weIl-known
color pattern consisting of about 8 oblique bands on the body, the 3rd
extending onto the ventral fio, the 4th on the dorsal fin and the 6th on the
anal fin.Ao almost identicalcolor pattern is seen in Leporinus fasciatus,
Lepon'nus desmotes and Synaptolaemus cingulatus.
There was a revision of the genus in the twenties, but it is now out of
date. The following tentative key wiIl serve to distinguish the few forms
recognized by the author:
a. Anal fin with 13or 14branched rays; 6!/41-42/5 scales; 4 transverse bands, their
direction perpendicular to the long axis of the body
. . . A.eques (species restricted
to the Rio Magdalena in
Colombia)
aa, Anal fin with 11 or 12 branched rays; usual1y 6/39-41/6 scales; 8 transverse
oblique bands . . . A. hypselonotus, a
superspecies or polytypic
species. Two forms may
eventual1y be distinguished:
-Abramites hypselonotus
ternetzi, restricted to the
Rio Paraguay basin.
Possible synonyms are
Leporinus solarii and A.
nigripinnis. It is said to have
the dorsal fin farther back
and the caudal peduncle
deeper than the Amazonian
formo
-Abramites hypselonotus
hypselonotus, the common
form from the Amazon
and Orinoco basins. A
probablesynonymis
Abramites microcephalus,
which cannot be told apart
from A. hypselonotus (by
usual taxonomic methods
at least). The young fish are
distinctly more elongate than
the adults.
175
Leporinus areus. Photo by Or. Herbert R. Axelrod.
Leporinus striatus. Photo by Or. Herbert R. Axelrod.
Leporinus sp., possibly L. nigrotaeniatus. Photo by Klaus Paysan.
- ---
t, ,
I
.......-
Abramites
hypselonotus.
Photo by Harald
Schultz.
.
"
t
A young Abramites hypselonotus. Photo by Or. Herbert R. Axelrod.
t""
.
,. "
~
.
y \'
..
~
*'
"""
""'-
.,
~
"
, ~
..
.
~.
;f'/"
;# ,
"'ri
I
.,
" ~
"
W~ ~ !
.
GENUS RHYTIODUS
This peculiar genus inc1udes the most elongate anostomids (depth up to
7 in the standard length); they have tinely denticu1ated teeth, usually 4 above
and 3 or 4 below on each side, their cutting edges respectively convex and
concave, strongly compressed from front to behind and relatively thin, as
compared to the robust teeth of most leporins. The species are of rather
large size (300 mm in standard length is not rare), and live in large rivers
(pelagic forms); they are of variable color, usually yellowish with dark
mottlings, or the reverse. They may be identified as follows (Rhytiodus
e/ongatus, which is curiously the least elongate, was not seen by the author):
a. 48-56 lateralline scales, 7-8/4-5 transverse sca1es.
b. Depth 4.8 in standard length; 48 lateralline scales; teeth said to be 3/3
. . . R. elongatus(RioPurus)
bb. Depth 5.2-7.0 in standard length; 52-56 lateralline scales; teeth 4/4
. . . R.argenteofuscus (Amazon
basin) (including "variety"
unifasciatus?)
aa. 76-90 lateralline scales, 11-12/8-9 transverse scales (depth 5.5-6.2 in standard
length; teeth 4/3 or 4) . . . R.microlepis (Amazon
basin)
GENUS ANOSTOMOIDES
This genus inc1udes two deep bodied species which have teeth like
certain leporins (somewhat trifid above, nevertheless) and the mouth
upturned like in the subgenus Laemolyla. These two species, Anostomoides
atrianalis from Venezuela and Anostomoideslaticeps (first described as a
Schizodontopsis,which is synonymous with Laemolyta) from the Essequibo
River in Guyana, are very c10seto each other. They differ in the following
way:
a. No marksactuallyvisibleon body; anal fin with 10branchedrays
. . . A. atrianalis
aa. 3 transverse bands plus some darkening of caudal base; anal fin with 8 branched
rays ... A.~nups
GENUS ANOSTOMUS
LISAS AND HEADSTANDERS
This important genus, the second (in number of species) after Leporinus,
has the body elongate and the teeth multicuspid (at least above) like
Schizodon, but the mouth c1early upturned. It inc1udes two groups of
species which can be considered as distinct genera or subgenera, a matter
of opinion. These two groups, which will be treated here as subgenera
because oftheir common characters, are Laemo/yta and Anostomus(typical).
178
"""!""""
4 ~
.........
Anostomoldes latlceps (after Elgenmann, 1912).
Laemolyta species are usually of a rather large size, but this is also the case
of one species of Anostomus; they usually have the mouth slightly less
upturned (not quite on the top of the head), but this character is somewhat
subjective; their teeth are c1assically multicuspid above and truncate below,
but the lower cusps disappear only at a certain age (corresponding to 60 or
70 mm in standard length), and the distinction between a young Laemolyta
and a young Anostomus is quite difficu1t and often one must guess; finally
some Laemolyta, at least, frequently adopt a head-down position as do most
species of Anostomus. As concems the geographical distribution, the more
generalized Laemolyta species inhabit, as a rule, the center of the Amazon
basin as well as the middle Orinoco, with some exceptions, whereas the more
specialized Anostomus species (together with Synaptolaemus, Sartor and
/ Gnathodolus) are more peripheral, inhabiting mostly the Guianas and the
south of the basin, again with exceptions.
As with Leporinus, the Laemolyta-Anostomus species will be characterized
in the shortest possible way, i.e. by means of identification keys.
Key to the subgenus Laemolyta (lisas):
a. A well-delimited longitudinal band; back with longitudinal lines; less than 6
scales between dorsal fin and lateralline (body compressed, the thickness 1.5-1.8
in the depth).
b. Transverse scales4/4
bb. Transverse scales 5/4 or 5
. . . A. (L.) garmani (Central
and Upper Amazon basin;
also Upper Rio Xingu)
. . . A. (L.) taeniatus (Central
Amazon basin; "Commonin
the Rio Negro)
179
Rhytiodus micro/epis (above) and Rhytiodus argenteofuscus (below)
from Rio Solimoes. Photo by Dr. Herbert R. Axelrod.
Unldentlfled young anostomid. Photo by Klaus Paysan.
f
~
A group of anostomid fishes. Photo by Klaus Paysan.
Adult Anostomus (Laemo/yta) proximus. Photo by Dr. Herbert R.
Axelrod.
t
I
...
,.
aa. Longitudinal band usually difi'use or broken; back with transverse bands, at least
in young specimens; more than 5 scales between dorsal fin and lateralline.
c. Transverse scales 6/5 or 6 (proximus-varius-nitens complex).
d. Longitudinal band not well delimited but continuous; depth 4.25-4.7 in
standardlength . . . A. (L.)proximus, possible
synonym A. varius
(sympatric with A. taeniatus)
dd. Longitudinal band broken; depth 5 in standard length
. . . A. (L.) nitens (sympatric
with A. taeniatus, also
present in Lower Amazon
basin)
cc. Transverse scales 7 to 9/6 to 7 (50 to 55 lateralline scales).
e. Longitudinal band extends to end of caudal fin; depth 4 and head 4.66
in standard length; thickness of body less than 1.8 in depth (?)
.. . A. (L.) orinocensis
(Orinoco)
ee. Longitudinal band not extending onto caudal fin; depth 4.2 and head
4.25 in standard length; thickness of body more than 1.8 in depth.
f. 9/50/7scales . . . A. (L.) fernondezi (Upper
Orinoco)
ff. 7 or 7;/54-55/6 or 6; scales . . . A. (L.)petiti(Rio
Araguaia)
Key to the subgenus Anostomus (headstanders):
In this group of species, the teeth are bi- or tricuspid in both jaws and the
mouth is exactly superior, that is the lower jaw is vertical when the mouth is
closed. As with the other genera, the species will be- defined by a key:
a. 42-46 lateralline scales (dorsal fin far in advance of midbody; color pattern of
lateral spots and longitudinallines through the center of the scales).
b. Usually a large spot on flank under dorsal fin, and a small peduncular spot;
several bars astride back; depth 3.2-3.9 in standard length in adult; dorsal fin
usually midwaybetween snout and adip~se fin; scales 5 or 6/42-44/4; or 5,11;
to 13in predorsal series; about 120-140mm ... A. (A.) trimaculatus(the
typical form, fromMato
Grosso, is perhaps different
from the specimens collected
in the Amazon, Rio Negro,
Guyana and Venezuela. It'
has the depth 3.2-3.4 in
standard length, as opposed
to 3.6-3.9for the" ..
Amazonian-Guianan
form) ,
bb. Usually 4 rounded spots on flank; no dorsal bars; depth 4.15-4.20in standard
length in adult; dorsal fin between snout and point behind adipose fin; 5 or 5;
/44-46/4t or 5 scales, 13-14 in predorsal series; about 160-180mm
. . . A. (A.) gracilis(Rio
Guapore; Rio Negro;
Orinoco)
182
..,....
Anostomus spllocllstron. Photo from the Zoologlcal Museumof Am-
sterdam,courtesy of Dr. H. Nljssen.
aa. 38-42 lateralline scales.
c. 4 roundish or slighdy horizontally elongate spots on fiank; longitudinallines
between rows of scales; about 12bars across back, some extending onto fiank
between the spots . .. A.(A.)plicatusandA. (A.)
spiloclistron(Guyana
and Surnam)
cc. Body with large longitudinal bands on fiank.
d. Color pattem formed of 2 narrow light longitudinal bands on a dark
background; back striated; dorsal fin usually at midbody except in very
large specimens; 12 or 13 predorsal scales (pectoral fins i, 14-15),
. . . A. (A.) ternetzi (Guianas;
Orinoco; Rio Araguaiaand
Rio Xingu)
dd. Color pattern composed of 3 dark longitudinal bands on light background;
back uniforrnlybrown, not striated; dorsal finusually in front of midbody;
11 predorsal scales (pectoral fins i, 12-14).
e. Depth 3.33-3.75 and thickness of body 8-13 in standard length;
transverse scales 5 or 5;/4 or 4t; snout relatively long, the eye slighdy
/ behind middle of head (snout-to-dorsal1.05-1.20 in dorsal-to-caudal)
. . . A. (A.) brevior (Mana and
Camopi Rivers in French
Guiana)
ee. Depth more than 3.8 and thickness ofbody 7.0-7.5 in standard length;
" transverse scales 4 or 4t/3; or 4; eye in middle ofhead (snout-to-dorsal o'
1.10-1.25in dorsal-to-caudal) . . . A.(A.)anostomus
(an elongate subspecies,
called A. anostomus
longus, comes from the
upper Amazon; the typical
comes from the Guianas
(except French Guiana ?),
Orinoco and (?) Colombia)
183
I
-----
-
Anostomus (Laemo/yta) taen/atus. Photo by Harald Schultz.
.. "'li~
~-
:f.II:oi..
..
-=:...
A young Anostomus (Laemo/yta)prox/mus. Photo by
Harald Schultz.
Lisa of the var/us-complex, possibly Anostomus (Laemo/yta) n/tens.
Photo by Dr. Herbert R. Axe!rod.
.....
Anostomus (Laemo/yta) petiti(or close to it). Photo by Harald
Schultz.
Unidentified lisas, probably close to Anostomus (Laemo/yta) taenla-
tus. Photo by Dr. Herbert R. Axelrod.
GENERA SYNAPTOLAEMUS, SARTOR AND GNATHODOLUS
SPECIALlZED HEADSTANDERS
These genera, each with only one species, are modified offshoots of an
Anostomus-like ancestor or similar formoThey are confined to very special
biotopes, such as crevices in rocks or faIlen trees, and they inhabit
restricted areas (having been coIlected, up to now, only in the "bifurcation"
of the Upper Orinoco and the Rio Casiquiare and in the Upper Xingu).
Synaptolaemus cingulatus(from both localities) has a striking coloration
composed of seven narrow yeIlowto orange rings encircIing the body on an
almost black background; as in Abramites and Leporinus desmotes, the
dorsal, ventral and anal fins are incIuded, in part, within the dark pattern.
Its teeth are the least specialized ofthe three: the median teeth ofboth jaws
are not enlarged and simply truncate, and the lateral ones are slightIy bi- or
trifid. It eats only insects, judging from the stomach contents.
Sartor respectus(from above the faIls in the Upper Xingu) has a quite
trivial color pattern (6 or 7 inconspicuous transverse dark bands and longi-
tudinal lines between the scales-rows) but a striking dentition, the lower
median pair of teeth being enlarged and pointed upward and backward
like a pair of tusks; likewise, the eyes are directed o'bliquely upward and
backward. It has a mixed diet, eating algae, insects and sponges.
Gnathodolus bidens. Holotype from the Canal de Casiqulare (after
Myers, 1950).
186
Gnathodolusbidens,from the bifurcation of the Orinoco and Casiquiare
rivers, also has a quite common coloration (two diffuse round spots and a few
bars on the back), but an even more striking dentition, the lower teeth having
disappeared except for the median pair, which also protrude out of the
papilIae of the lips, but with a sickle-shaped form; the eyes are directed
obliquely backward. Nothing is known of its diet.
Comparison of Synaptolemus with Anostomus and Gnathodolus. (a)
Rlght premaxlllary and right mandibular teeth; (b) positlon of teeth on
jaws; (c) side vlew of head (after Myers, 1950).
&
Synaptol aemus
w.,
Anostomus
~~
.,
~:. :-
. ..
a \)
J
f Ib
~
-
Synaptolaemus
,~
: i b
~
e
Q a
<::::>
Gnathodolus
~
187
.~~t
\' ,
\ ' '"
"
"
Anostomus (Anostomus) ternetzi. Photo by Harald Schultz.
Head of Anostomus ternetzi. Photo by Harald Schultz.
188
-.......-
I
Anostomus (Anostomus) anostomus. Photo by R. Zukal.
189
Family Hemiodidae
Hemiodids and their relatives
These fishes, of rather smalI size with the body usualIy fusiform and the
mouth subinferior, resemble somewhat the leporins and headstanders, but
with a very different dentition: in most of the species the lower jaw is
reduced and lacks teeth. On the upper jaw as welI as someimes on the feeble
maxilIary, the teeth are smalI, more or less cuspi date incisors which are
loosely inserted into the bone or even into the gums. We will find this
regressive tendency again in the chilodins and prochilodins which folIow,
whereas in the next (and last) step, the anodins and curimatins have
completeij lost their teeth, at least after the postlarval stage.
In the course of their evolution, the hemiodids have diverged rather
considerably, according to their respective ecology (surroundings). At least
three groups may be recognized: (1) species living in the rapids or open
waters; (2) species living on (in) the sandy bonom; (3) species living on
the bonom, usualIy on rocks, rarely on sand. They can be distinguished by
the folIowing characters:
a. Mouth subinferior, more or less protractile; no maxillary teeth; fontanels present;
adipose lid present; paired fin not very developed
. . . Subfamily HEMIODINAE
b. Gill membrane free from isthmus; scalescycloid; mouth not much protractiie;
fontanel broad; gill-rakers normal; buccal cavity without papillae; live in open
water.
c. Teeth unicuspid in both jaws; lower jaw with two rows of teeth in its anterior
part . . . Tribe Micromischodini
cc. Teeth multicuspid; no teeth in lower jaw after the post-larval stage
. . . Tribe Hemiodini
bb. Gill membrane partly united to isthmus; scales cyclo-ctenoid; mouth strongly
protractile; fontanel narrow; inner face of gill-rakers with numerous tubercles;
buccal cavity with several valvulaeor papillae; live on sand beaches (teeth very
weak, tricuspid) . . . Tribe Bivibranchiini
aa. Mouth strictly inferior, with the snout scarcely movable; maxillary and mandi-
bular teeth present but never numerous; fontanels closed; adipose lid absent;
paired fins developed; bottom-fishes (teeth usually broad incisors, truncate or
multicuspid; gill membrane not united to the isthmus; scales cycloid)
. . . Subfamily PARODONTINAE
SUBFAMILV HEMIODINAE
Tribe Micromischodini
The single species Micromischodussugillatusis one of the most interesting
discoveries of recent years. It was colIected near Manaus by Dr. Tyson
Roberts, who described it in detail in 1971.
190
r
..
Micromischodus sugillatus (after Roberts, 1971).
It is the only hemiodid with numerous minute teeth on both jaws, and
one of the few non-characids with two rows of teeth on the anterior part of
the lower jaw (about 55 to 60 in each row), yet in the shape of the body it
does not differ at alI from a more "conventional" hemiodin such as
Argonectes.
Tribe Hemiodini
Four genera comprise this group characterized by multicuspid teeth
present only on the premaxilIa (except at an early post-Iarval stage, where
some mandibular teeth are to be seen, as in Curimata), the snout scarcely
mobile except in Argonectes, and a few other technical characters to be seen
in the folIowing key to the genera:
a. Teeth with more than 3 cusps; adipose lid not entirely covering pupil; snout
scarcely protractile; gill membrane not united with isthmus.
b. Dorsal fin small; mouth usually subinferior, with the teeth almost horizontal,
backward directed and almost parallel to premaxillary bone.
c. Scales of back only a little smaller than those of abdominal region
. . . Hemiodopsis
cc. Scales of back much smaller than those of abdomen; ratio of apparent
height of scales below and above lateral line 1.8-2.2 in adult
. . . Hemiodus
bb. Dorsal fin filamentous, at least 1.5 times longer than head; mouth almost
terminal, with teeth almost vertical and inserted obliquely relative to premaxil-
laryaxis . . . Pterohemiodus
aa. Teeth with 3 cusps, loosely inserted into the lips; adipose lid completely covering
pupil; snout partIy movable; gill membrane united forward to isthmus (an almost
perfect link with the next tribe, the Bivibranchiini)
. .. Argonectes
191
li
a
h~*~1i _ -- -- --.
Anostomus (Anostomus)trmaculatus.Photo by Harald Schultz. . Synaptolaemus cngulatus from the Upper Xingu. Photo by Harald
Schultz.
Anostomus (Anostomus) plicatus. Photo by Harald Sch'ultz. Sartor respectus from the Upper Xingu. Photo by Harald Schultz.
-- ---
,
GENUS HEMIODOPSIS
With at least 15 species, Hemiodopsis(formerly known as Hemiodus,that
actually replaces Anisitsia) is the largest hemiodid genus. The followingkey
was published in Tropical Pish Hobbyist (November, 1964); it is here
somewhat modified: r
a. Less than 80 lateralline scales..
b. No definite transverse bands on sides; usually a longitUdinal band and/or some
spots.
c. Less than 55lateralline scales; transverse'scales 7 or 7t/4 or 5; a longitUdinal
band (gracilis-group).
d. Body slender; 42 lateral line seales . . . H. gracilis(Amazon,
Guyana and Rio Sao
Francisco)
dd. Body less elongate; 44 or more lateralline scales.
e. 44-46 lateralline scales.
f. LongitUdinal band beginning below dorsal fin levei
. . . H. goeldii, a species often
confused with H.
semitaeniatus (Guianas and
Rio Xingu)
ff. LongitUdinal band continued anteriorly by a lighter band to opercle
. . . H. thayeria (Rio Negro)
ee. 50-53lateralline seales(longitudinalband beginningbelowdorsalfin
leveI) . . . H.fowleri (Amazon)
ec. 55-72 lateralline seales (semitaeniatus-group).
g. A longit],ldinalband; less than 64 lateralline scales.
h. 9/56-60/5 seales.;longitudinal band beginning beIow dorsal fin
leveI . . . H.semitaeniatus (Rio
Guapore)
hh. 11/63/6scales; longitUdinalband continuous from eye to end of
lower caudallobe . . . H.ternetzi (RioToeantins)
gg. No longitudinal bando
i. No marks on fiank; 10-11/69-72/6or 7
. . . H. immaculatus (Rio Negro
and middle Amazon basin)
ii. A round spot on the middle of the fiank; 17-18 (?)/65-70/11
scales .. . H. rodolphoi (RioParahyba)
bb. Three transverse bands on 9:ideand one on caudal peduncle (quadrimaculatus-
group).
j. 42-45 lateralline scales.
k. 9 branehed anal rays; about 27 gill-rakers on lower arch
. . . H. quadrimaculatus
quadrimaculatus (Camopi
River in Freneh Guiana)
kk. 8 branehed anal rays; 16-21 gill-rakers on lower arch
. . . H. quadrimaculatus
vorderwinkleri (Upper
Amazon; Guyana,
Surinam)
194
r
"
~
k
(
f
~.
I.
I
I
I
jj. 50 or more lateralline scales.
1. 9t-l0/50,52/5 or 6 scales
. . . H. huraulti (Surinam and
French Guiana)
lI. 12/64-67/7t or 8 . . . H. sterni(Upper Rio
Juruena)
aa. 80-130 lateralline seales (usually a round spot on the fiank) (microlepis-group).
m. 20/80-83/12 scales
. . . H. parnaguae (Rio Pamagua;
a similar form collected in
Surinam, still undescribed,
has 81-99 lateral line scales)
mm. More than 100 lateralline scales.
n. 24-25/110-112/14-15 scales
. . . H. microlepis (Amazon, Rio
Negro, Rio Guapore; a
similar form collected in
Bolivia, still undescribed,
has 27-28/21-23 transverse
seales)
nn. 30-32/120-125/18 or 19 scales
. . . H. argenteus (Rio Orinoco)
Paratype of Hemiodopsis huraulti. Photo by the author.
195
Unldentlfled Hemiodopsis of the gracilis.,group. Photo by
Harald Schultz.
Hemiodus unimaculatus or Curimata ocel/ata, mlmetlc species from
the vlclnlty of Humaita, Brazll. Photo by Dr. Herbert R. Axelrod.
-
.-':11
4
~~
~ v,,-
,!~r.
~ ..
~
...
,,,..
, .
196
...
~
,
~.
(
Hemodopss gracilis. Photo by Hilmar Hansen.
Hemiodopss of the microlepis-group from Humalta, Brazll. Photo by
Dr. Herbert R. Axelrod.
r ,
~
,,~p.~
.
" .
..
. .', "
.
' ~.it...
.
IS'''
, '.. t~..
1rrI;;," ... ,'L .~~.1iI,$'1f(,r
~~ ,:J,~~~" '''''-~.''
J'i:L.o.,;",;.:;;
e. ~~ fi
~ ""!r.. ___,
~, ~;
,
~~::/
'--~'~"" ~
"
' ..<;
,
'"
,
'='" -,. "
"
-
.
----
197
GENUS HEMIODUS (synonym ANISITSIA)
In this small group the scales of the back of adult or semi-adult fishes
are notably smaller than those above the ventral fio; apart from this charac-
ter there is nothing which distinguishes Hemiodus from Hemiodopsis,and
old descriptions (where the size of the scales is not mentioned) are
unreliable in this respect.
Hemiodusunimaculatus,formerly called Anisitsia notata, is Sifusiform fish
strongly resembling a Hemiodopsisof the microlepis-groupfor example,
with a round black spot astride the lateralline behind the dorsal fin leveI.
It is.common in the Guianas, the Rio Orinoco and probably in the state of
Para (under the name H. microcephalus), and even in the Rio Guapore. To
the south it is apparently replaced by a species with more lateralline scales
(about 90 instead of 60-72), H. orthonops,from the Rio Paraguay.
GENUS PTEROHEMIODUS
The two known species, P. atranalis and P. luelingi, both from the
Upper Amazon, have the body shape and the color pattern of a generalized
hemiodin such as Hemiodusunimaculatus,but the dorsal fio very high and
the teeth obliquely set.
I
I
I
GENUS ARGONECTES
This genus, together with Atomaster, forms a transition between the
"true~' hemiodins such as Hemiodus, and the "true" Bivibranchiini.
Following recent studies, it seems that the two known species may be
united into a singleone, Argonecteslongiceps(older name than A. scapularis),
with a certain variability in the color pattern according to the localities
( Rio Negro, upper Rio Orinoco, Lower Amazon and the Maroni River
between Surinam and French Guiana). Specimens from the upper
Orinoco (the form called A. scapularis)have a black, collar-like marking
bordering the gill-opening, dark pigment on the dorsal fin, dark bands on
the caudal lobes, and an elongate band behind the dorsal fio leveI. In
spedmens from the Maroni River, the mark on the flank is less elongat'b
and resembles that of Hemiodusunimaculatus.
Tribe Bivibranchiini
The members of this group are hemiodins having adapted themselves
to a diet composed at least partly of small animaIs living in the sand. They
have thus developed a strongly protractile mouth to suck in the sand and
bucco-pharyngeal papillae and valvulae to filter it. Two genera are known,
198
......
1/
u
r1
-
-011:
~~
~""\
1
I h ,YM tn ,/(1;. i<AI>!~'
<.,~",,' .
;;: "
.....
Holotype of Pterohem;odus atranalis (after Fowler, 1950) and B;v;-
branch;a protractila (after Elgenmann, 1912).
Atomaster, which is still rather dose to the hemiodins in the strict sense, and
Bivibranchia, which is a very specialized formo The two genera differ in the
following way:
a. Gill membrane free from isthmus; sca1es very numerous, their borders pectinate;
valvulae on roof of mouth not much developed .,. Atomaster
aa. Gill membrane united to isthmus (on a small surface); scales not more than 62,
crenulate only in adult; valvulae on roof of mouth well developed
. . . Bivibranchia
The single species of Atomaster, A. velox, was collected in 1924 in the Rio
Tocantins, state of Goyaz, Brazil. It has not been found again. It is straw
colored with a longitudinal silvery bando
Bivibranchia protractila exists not only in the Rio Tocantins, but also
along the sand beaches of the Branco and Negro rivers, near Santarem, and
in Guyana. It has a plain coloro Two other species have recently been found
in collections from Surinam.
199
Hemiodopsis goeldii. Photo by Harald Schultz.
Hemiodopsis of the gracilis-group. Photo by Dr. Herbert R. Axelrod at
Rio Negro.
200
:f
Hemiodopsis sterni. Photo by Harald Schultz.
Hemiodopsis quadrimaculatus vorderwinkleri. Photo by Dr. Herbert R.
Axelrod.
SUBFAMILV PARODONTINAE
The members ofthis group (that deservesfuII famlial status according to
certain ichthyologists) are all specializedfor life on the river bottoms, mostly
on rocky soi!: the mouth is strictly inferior, the breast flat and broad, and
the paired fins, chieflythe pectorals, are well developed and used to "walk"
on the bottom. Other more technical characters have been given in an'
earlier key. Two genera are here recognized, as follows:
a. 80r 10multicuspidincisiorteethon a straightlinealongthe premaxillarybones;
usually2 similarmaxillaryteeth; circumorbitalbonesnot coveringface
.' . . Parodon
aa. 6 truncate incisior teeth, not in a straight line on the premaxillary bones; no
maxillary teeth; circumorbital bones well developed, covering practically the
whole face . .. Saccodon
GENUS PARODON
Within this rather important unit, a few subunits may be found. They
will here be called subgenera and may be distinguished as follows:
a. Pectoral fin with only one unbranched ray; ventral fin usually wi~ i, 7 rays.
b. 1 to 3 small truncate teeth on sides of lower jaw; upper teeth usually with a
straight edge . . . Subgenus Parodon typical
bb. Lower jaw not toothed in adults; upper teeth usually with a rounded edge
. . . Subgenus Apareiodon
aa. Pectoralfin withtwo unbranchedrays;ventralfin with i, 8 rays
. . . Subgenus Parodontops
Subgenus Parodon
The species of this subunit are perhaps the least specialized of the Paro-
dontinae. Two speciesaresomewhat independent from the rather generalized
type. The first, for which Fowler proposed the generic name Nematoparodon,
is Parodonapolinari,from the upper course of the Rio Orinoco in Colombia
and Venezuela. It is remarkable in having a filamentousdorsal fin and
tubercles on the snout. These peculiarities are generally considered as
secondary sexual characters, and are much less common in the characoids
than in the cyprinoids for example. Parodon apolinari is convergent to a
Labeoin this respect. The other species is Parodonguyanensisfrpm French
Guiana and Surinam, which has rounded indsor teeth, as in Apareiodon,
numbering 10instead of8 as in other species ofthe subgenus; it also has an
elongate body. Apart from these characters which are probably of sub-
generic rank, P. guyanensisis recognizableat first sight by a black bar on the
dorsal fin and a zigzag mark along the lateral line.
The other 9 species, very close to each other, may be distinguished by
means of the followingkey (aduIts only):
a. More than 20 cuspson middleupper teeth in adult; an irregularlongitudinal
band. . . . P.buckleyi (Ecuador)
aa. Lessthan 20 cuspson middleupper teeth in adulto
202
--------
;t
~
~
,.
,
L
b. More than 12 cusps on middle upper teeth.
c. Depth usually 3.5-3.8 in standard length.
d. A series of vertical dark bars on sides; anal fin ii, 6; scales 5/39 or 40/4 or 5,
13! or 14! in a predorsal series . . . P. caliensis(Colombia; an
Andean species living at
1000-1100 m altitude)
dd. Usually one or two longitudinal bands; anal fin ii, 7; scales 4 or 4t/
35-39/3 or 4, lI! or 12t in a predorsal series.
e. Scales 4/35-38/4; about 17-18 cusps on middle teeth; caudal lobes
equal; head about 4.35 in standard length; one longitudinal band
. . . P. suborbitale, type of the
genus (Venezuela; perhaps
Colombia)
ee. Scales 4/39/3; about 15 cusps on middle teeth; lower caudallobe more
elongate than upper one; head about 5.15 in standard length; two
longitudinal bands . . . P. bifasciatus,a very rare
species, apparently slightly
differeI1t:from suborbitale
(Rio Branco basin)
cc. Depth usually more than 3.8 in standard length.
f. Depth 3.9-4.25 in standard length; pectoral fin usually i, 12-13.
g. A longitudinal band up to end of middle caudr.1 rays; anal fin ii, 6
. . . P. carrikeri (upper Rio
Paraguay)
gg. A series of about 15 lozenge-shaped spots along lateralline; anal
fin ii, 7 . . . P.gestri(MatoGrosso)
ff. Depth 4.2-5.6 in standard length; pectoral fin i, 14 or 15
. . . P. tortuosus (probably 2
subspecies, P. t. caudalis
and P. t. tortuosus,
sympatric, respectively,
with P. carrikeri and P.
gestri and having,
respectively, the same ..
pattem (sibling forms ?)
bb. 9-12 cusps on middle upper teeth.
h. A very conspicuous, well delimited black longitudinal band
(depth 3.8-4.6 in standard length)
. . . P. pongoense, originally
described as an Apareiodon,
synonym A. caquetae
(Peruvian Amazon,
Ecuador and Colombia)
hh. A diffuse dark longitudinal band (depth 4.5-4.66 in standard
length) . . . P. nasus(probably2
subspecies, P. n. nasus from
Mato Grosso and P. n.
hilarii from Minas Gerais,
differing in the number of
lower teeth and other details)
.. (P. pirassunungaefrom S.E. Brazil is a synonym of P. tortuosus)
203
Bivibranchia sp. Photo by J. Elias.
A Hemiodopsis of the microlepis-group. Photo by R. Zukal.
Parodonpongoense. Photo by Or. Herbert R. Axelrod.
Parodon sp. close to P. affinis. Photo by Klaus Paysan.
Bivibranchia sp. Photo by J. Elias.
A Hemiodopsis of the microlepis-group. Photo by R. Zukal.
Parodonpongoense. Photo by Or. Herbert R. Axelrod.
Parodon sp. close to P. affinis. Photo by Klaus Paysan.
SubgenusApareiodon
The species of this group, 8 or so in number and all from eastern and
southeastern Brazil and the La Plata basin, curiously parallel the Parodon
species in the evolution of their meristic characters and coloration. It is not
impossible that, in the future, they will be distributed differentIy within the
genus because the presence or absence of a few mandibular teeth as well as
the more or less rounded edge of the upper teeth seem to be rather poor
characters, at least at the generic or subgeneric leveI.
The following key is partIy compiled from the literature and it is by no
means intended as a revision of the subgenus:
a. Usually two maxillary teeth on each side (occasionallyone in P. piracicabaeand
P. davisi).
b. Rather deep bodied species, depth usually less than 4 in standard length
. . . P. davisi (N.E. Brazil);
bb. Elongate species, depth more than 4 in standard length.
c. 35-36 lateralline scales and 10-11predorsal scales; usually ii, 9 dorsal rays
(a conspicuous black longitudinal band plus a dorsal band)
. . . P. irapicuruensis (Rio
Itapicuru, coastal river of
Eastern Brazil near Rio S.
Francisco basin)
cc. 39-441ateralline scales and 12-14 predorsal scales; usually ii, 10 dorsal rays.
d. Dorsal fin with 2 black bars; 39 lateral line scales
. . . P. machrisi (near Rio de
Janeiro)
dd. Dorsal fin plain; 40-44 lateralline scales.
e. A conspicuous longitudinal band plus another one on back
. . . P. piracicabae (basin of the
Tiete-Parana rivers, S.E.
Brazil )
ee. A longitudinal line plus 6 transverse bands on back (usually less
conspicuous) . . . P. affinis, synonymP.
paraguayensis (La Plata,
Parana, Paraguay basins)
aa. Only one maxillary tooth on each side (depth 4.33-5.0 in standard length).
f. Less than 40 lateralline scales (34-39), 14 predorsal scales; a Jongi-
tudinal band with 7-9 lozenge-shaped spots superimposed on it
. . . P. ibiriensis (S.E. Brazil)
ff. Usually 40 or more lateralline scales and only 10-12 predorsal scales.
g. Head 4.5-5.0 in standard length; 4 or 4t/3 transverse scales; a
longitudinalline along body . . . P. hasemani (RioS.
Francisco)
gg. Head 4.4 in standard length; transverse scales said to be 4[2; a
series of bars or lozenge-shaped spots along body
. . . P. mogiguacuensis (Rio
Mogiguacu)
206
~ l
firti'..
t
Saccodon wagneri, showing ventral view of head and thorax and teeth
(after Eigenmann, 1922), and Saccodon dariensis (after Meek & HiI-
debrand, 1916).
(}
.
Subgenus Parodontops
The single species Parodon ecuadoriensis(synonym Parodon terminIJlis)
from the Pacific Coast of Ecuador represents one of the extremes of the
specializationtoward a bottom lifeand strongly parallels certain Characidium.
It has the breast broad and fiat and the pectoral fins very low, horizontal,
and spread like those of an Eleotris (see footnote under Saccodon). Apart
from the usual irregular longitudinal dlJrkband, this species has an orange
caudal fin barred with black.
,
GENUS SACCODON
With their armor-plated head, species of the genus Saccodon are strange
fishes indeed, looking like remnants of an older age now surviving in some
remote torrents of the Andes and Panama. According to Schultz and Miles,
the species can be identified in the following way:
a. Pectoral rays ii, 15-16; 39-41 lateralline scales.
b. Dorsal rays ii, 10; anal rays ii, 7; color pattern of elongate dark blotches
arranged in three streaks along sides; caudal and anal fins spotted
. . . S.caucae (Upper Rio
Cauca)
bb. Dorsal rays ii, 9; anal rays ii, 8; color plain; fins not spotted
. . . S. wagneri(westernslope
ofEcuador)
aa. Pectoralrays ii, 12; 35-37lateral line scales(color pattern much like that of
S.caucae) . . . S.dariensis (Panama)
..
4-
207
Parodon affinis (or close to it). Photo by Harald Schultz.
Parodon suborbitale. Photo by Dr. Fernandez-Yepez.
208
t
Two aspects of Caenotropus labyrinthicus. Upper photo by Harald
SChultz, lower photo of a specimen from Puerto Gaitan, Colombia by
Dr. Herbert R. Axelrod.
i9
In an interesting paper just published, T. Roberts (1974) suggests that eertain
northwestem Parodontinae, so far separated into distinct species or even genera on
the basis of the structure of their teeth (as it was done also here), are aetually variant
individuals ("morphs") of a Saccodon.These would be S. dariensisin Panama and
Colombia (synonym S. caucae,aeeording to Roberts), and S. wagneriin Ecuador. In
other words the genus Saccodonwould present a "dental polymorphism", having the
teeth variable in form as well as in number.
Following this interpretation (of eourse to be tested against other hypotheses such
as mimiery, sibling speciesand so on, by means of erossing experiments and/or study
of the ehromosomes, for example), P. (Parodontops) ecuadoriensis(see preeeding
section) would fali int the synonymy of Saccodonwagneri and Parodontops(type
ecuadoriensis) thus would beeome a synonym of Saccodon.
Family Curimatidae
Curimatas
With the curimatids we reach the most regressed of the characoids, in a
certain sense. Most of the curimatids have no teeth when adult, and they
content themseIveswith food found in the mud, i.e. mostly algaeand detritus
like many carps; others have a mouth in the form of a sucking disc, with
which they,rasp their food from diverse substrata. All have the long intestine
of typical herbivorous animals.
Four groups are here recognized within the curimatids, some of which
are considered as distinct families by certain ichthyologists.* They can be
diagnosed as follows:
a. Anal fin with 3 or 4 anterior unbranehed rays; fourth branehial arch thiekened;
swimming position head down (7-10 branehed dorsal rays; teeth present, few in
number,bifid,setintothelips) ... Subfarnily CHILODlNAE
aa. Analfin with 2anterior unbranehed rays; fourth branehial areh normal; swimming
position horizontal (S or more rarely 9 branehed dorsal rays).
b. Mouth protractile, forming a sueking dise, with two eomb-like rows of teeth
set into the lips; a predorsal spine (gill membranc united to the isthmus; body
brcam-shaped; mosdy detritivorous) . . . Subfamily PROCHILODINAE
bb. Mouth scareelyprotractile, not forming a sueking dise, lacking teeth after the
post-larval stage; no predorsal spine.
c. Bottom fishes, with the body shape of a bream, very similar to that of a
prochilodin, the breast and belly often flattened or, on the contrary, kee:d;
gill-rakers usually rudimentary, gill membranes united to the isthmus;
adipose lid usually absent; maxilla redueed; mosdy mud eaters when adulto
. . . Subfamily CURIMATINAE
* T. Roberts (1973) presents some osteological evidenees favoring relationships
between Proehilodinae (whieh he treats as a farnily and ealls Proehilodontidae, eon-
trarily to the Rules) and Anostomidae, rather than Curimatidae. However the jaws,
the teeth and the general organization of the two first named groups are markedly
ditIerent.
210
.
,
ec. Pelagic fishes with the body fusiform, very similar to that of a hemiodin, the
breast and belly rounded; gill-rakers long and numerous,gillmemb"ranesfree
from the isthmus; adiposelid eoveringthe eye; maxilla relativelydeve1oped;
plankton eaters . . . Subfarnily ANODINAB
SUBFAMILV CHILODINAE
Headstanders
The chilodins share with certain anostomids the faculty of swimming
head down: this must be in connection with some special ecological needs
and not due to dose relationships. Phylogenetic connections are more likely
to be looked for in the direction of the hemiodids (according to certain
iChthyologists) or the curimatids.
The author published some years ago (Tropical Fish Hobbyist, May,
1964, pages 5-10 and 63-67) a simplified review of the headstanders.
Although its exaggerated simplicity may be confusing, it is repeated below,
no one having since added essential data.
The two genera here recognized may be distinguished as follows:
a. Mouth distinctly inferior, with thin lower lip, the lower jaw included; pharyngeal
teeth rather large and relatively not numerous; scales erenulate in the adults;
predorsal region rounded exeept just before dorsal fin; post-ventral region
rounded; anal rays iii, 7i or'S, the fin with a eoncave border; rather large size
(up to IS5 mm) . . . Caenotropus
aa. Mouth terminal or slightly upturned, the lips thiek, the.lower jaw slighdy pro-
minent when the mouth is closed; pharyngeal teeth rather small and numerous;
scalescycloid; predorsal and postventral regionskeeled; anal rays iii, 10or 10i, the
fin with a straight border; usually not more than 120mm in length
. . . Chilodus
.
GENUS CAENOTROPUS
HEADSTANDERS
Two species are known: Caenotropuslabyrinthicus and C. maculosus.
Caenotropuslabyrinthicus,from the Amazon basin, the Rio Orinoco and also
the Rio Parnaiba, has a black longitudinal band along its body, extending
from the tip of the snout to the end of the middle caudal rays, with a bulge
shonly before the dorsallevel; the first dorsal rays are dark. In the Rupununi
River there is a dose form characterized by a broken lateral band and a
conspicuous bar on the dorsal fino
Caenotropusmaculosus, from the Guianas, has about the same color
pattern, with the tip of the dorsal fin black. It differs in having fewer
longitudinal scales, postdorsal scales, gill-rakers and branched caudal rays,
a longer head, the adipose fin above the leveIof the last anal rays (instead
of above the first ones as in C. labyrinthicus),and some teeth on the lower
lip (instead of none or one or two very rudimentary ones as in C. labyrinthi-
cus). The followingtable gives the principal characters ofboth species:
211
Caenotropus maculosus. Photo by Dr. Herbert R. Axelrod.
Chilodus punctatus, juvenile. Photo by Dr. Herbert R. Axelrod.
- ----
212
A pair (?) of Chilodus punctatus. Photo by H.J. Richter.
213
GENUS CHILODUS
SPOTTED HEADSTANDER
The single known species (at least the single one regularly discussed),
Chiloduspunctatus, from the Guianas, upper Amazon and upper Orinoco
Rivers, is identifiable at once by the form of its cyprinoid mouth, somewhat
resembling that of a Pekinese dog, its higher body, its spotted dorsal fin,
and a different :::olorpattern of the body; the longitudinal line, con-
spicuous from snout to opercle, is only "suggested" posteriorly by a series
of brown spots at the center of each scale. During spawning (which is a
very rare event in captivity), the spotted pattern disappears and is replaced
by a broad humeral spot.
The spotted headstander is a rather variable species as regards the body
shape and color pattern. This is why Chilodus zunevei, from the upper
Maroni in French Guiana (whose types are lost), has to be considered, until
it is collected again, as particularly deep bodied specimens of the spotted
headstander (body depth about 2.55 in the standard length, versus at least
2.75 in the "normal" headstander). On the other hand, there are
indications that an undescribeq headstander does live in the upper Amazon
basin, more precisely in the Rio Japura. This species has the upturned
mouth, the long anal fin and spotted dorsal fin of Chiloduspunctatus, but
the elongate body and the longitudinal band of Caenotropus (either
labyrinthicusor maculosus).
SUBFAMILY PROCHILODINAE
Bocachicos
This is a rather important group of no less than 25 species, alI reaching a
fairly large size (30 to 45 cm). The biology of some, which are of
economic importance in Brazil or Colombia, has been rather extensively
studied by local ichthyologists, and the group has been recently reviewed
from a systematic point of view by Dr. Mago, with an accent on the
214
Venezuelan species. According to this review, three genera can be defined
as follows:
a. Predorsal spine forked in front; edge of labial teeth rounded; snout moderate in
length, jaws equal or upper one not very prominent, the mouth not inferior.
b. Scales cydoid or crenulate, the postdorsal ones in a regular series from dorsal
to adipose fin, finely striated; caudal fin with conspicuous bars
. . . Semaprochilodus
bb. Scales ctenoid, at least in adults, the postdorsal ones usually not in a regular
series and not striated, similar to those of ftanks; caudal fin plain or spotted,
the spots sometimes arranged in wavy lines but never forming conspicuous
bars . . . Prochilodus
aa. Predorsal spine blunt in front and not forked; labial teeth pointed; snout long,
upper jaw prominent with the mouth inferior (scales cydoid; caudal fin plain)
. . . Ichthyoelephas
.
GENUS SEMAPROCHILODUS
BOCACHICOS
The genus Semaprochilodus (possibly only a subgenus of Prochilodus) is
better characterized by the very peculiar color pattern of the caudal and
anal fins of alI species than by the strUcture of the sca1es. They can be
identified as follows:
a. Elongate species, body depth 2.8-3.3 in standard length (scales crenulate, 10 or
more in a transverse series from dorsal fin to lateralline; 3 or 4 oblique bars on
each caudallobe).
b. 66-76 lateralline scales and 12-13/12-14 or 15 (up to ventral fin) transverse
scales; body depth about 2.8-2.9 in standard length
. . . S. taeniurus (middle
Amazonbasin)
bb. 48-50 lateralline scales and 10/10 transverse scales; body depth 3.0-3.33 in
standard length . . . S. binotatus(RioBranco;
a dubious species)
aa. Deep species, body depth 2.0-2.7 in standard length.
c. 7!-9 scales in a transverse series above lateralline; scales with border feebly
crenulate.
d. 3-4 oblique bars on each caudallobe; 45-49 lateralline scales
. . . S. kneri (Rio Orinoco)
dd. 4 (juvenile)-7 oblique bars on each caudallobe; 38-41lateralline scales
. . . S. insignis (Rio Branco;
Guianas)
cc. 9-13 scales in a transverse series above lateral line.
e. 7-9 bars on each caudallobe in adult, rather indistinct in old specimens;
edge of scales scarcely crenulate (body depth 2.0-2.6; 47-57 lateralline
scales, 10-lI/I0-ll! transverse scales)
. . . S. laticeps (Rio Orinoco;
sympatric with S. kneri)
ee. 3-4 bars on each caudallobe in adults; edge of scales distincdy crenulate.
f. 47-52 lateral line scales, 9-10/9-10 transverse scales (body depth
215
labyrinthicus maculosus
Standard lengthJlength ofhead
3.6-3.8 3.4-3.5
Lateralline scales
(counting alI)
29-31 26-27
Postdorsal scales
(up to adipose fin)
9-10
6t or 7
Gill-rakers 19-23 14-16
Branched caudal rays
17 15-16
Unidentified Chilodus from Rio Japura in Brazil. Photo by Harald
Schultz.
216
Semaprochilodus taeniurus. Photo by Dr. Herbert R. Axelrod.
,Semaprochilodus theraponura from Rio Urubu. Photo by Harald
'Schultz.
.
.-
2.35-2.7 in standard length, possibly 2.8-2.85 in verylargeindividuals,
or in certain populations) . . . S.theraponura, synonym
S. amazonensis (middle and
upper Amazon)
ff. 54-61 lateralline scales, 12-13/12t-13 transverse scales (body depth
2.0-2.4in standard length) . . . S. squamilentus(eastem
Brazil)
GENUS PROCHILODUS
BOCACHICOS
The bocachicos ("smalI mouths") of the genus Prochilodus are among
the least variable of alI characoids. They lack typical characters permitting
their sure identification, and the present state of their systematics is poor.
The color pattem oftheir caudallP1d dorsal fins, as welI as their scale formu-
las, seem the only actualIy reliable characters, if associated with the body
depth (provided that it is measured on a good series of adult specimens).
The folIowing is a very crude estimate of what may be the real situation.
Two artificial groups may be found within the genus, depending on the
coloration of the caudal fin, which may be spotted (the spots somewhat
arranged in wavy lines, but never forming conspicuous bars as in
Semaprochilodus), or plain, often quite dark, without any pattem.
The first group (with the caudal spotted) may be calIed the nigricans-
group (from the oldest species name). Mostly according to the scarce data
of the literature, the species can be identified as folIows:
Key to the Prochilodus nigricans-group (caudal fin spotted):
a. Usually less than 50 lateral line scales.
b. Body depth 2.45-2.8 in standard length.
c. 43-46 lateralline scales and 7-9/5t-7 transverse scales
. . . P. rubrotaeniatus(Guianas,
Venezuela,Rio Negro and
Rio Branco, upper Amazon)
cc. 48 lateralline scales and 11/7 transverse scales
. . . P., beni (Bolivia)
bb. Body depth 2.8-3.5 in standard length.
d. Snout moderate in length, upper jaw not projecting (body depth 2.8-3.2
in standard length).
e. 37-43 lateralline scales (8/7 transverse scales)
. . . P. corimbata, possible
synonyms P. oligolepis and
P. cearaensis (eastem and
southeastem Brazil)
ee. 44-49 lateralline scales (8-10/7-8 transverse scales)
. . . P. nigricans and P.
ortonianus (two sibling
species (?) from the
Amazon; P. cephalotes may
be the young of one of
these)
218
T
.
dd. Snout elongate, upper jaw projecting, resembling that of genus
Ichthyoelephas.
f. Bodydepth2.9-3.0in standardlength; 40-43lateralline scalesand
6t-7t/5t transverse scales . . . P.maripicru (Guianas)
ff. Body depth 3.5 in standard length; 47 lateralline scales and 8/7-8
transversescales . . . P. caudifasciatus (Rio
Perene in Peru, a poorly
known species)
aa. 49-65 lateralline scales (body depth 2.3-3.1 in standard length and 10-12/8-10
transverse scales) . . . P. mariae(RioOrinoco;a
species very dose to
P. brama)
The second group, with the caudal fin plain (Prochilodus typical or
P. argenteus-group), is even less welI known than the preceding. Certain
species, such as Prochilodus vimboides with P. steindachneri and P. hartii,
Prochilodus reticulatus with P. magdalenae, Prochilodus lineatus with
P. affinis, etc., form groups, usualIy sympatrics, which can be sibling species
or simply synonyms. AlI species except P. reticulatus (with P. magdalenae)
are from eastem and southeastem Brazil, some extending south to
Argentna.
The folIowing key, like the preceding one, is mostly compiled from the
literature.
Key to the Prochilodus argenteus-group (caudal fin plain):
a. Dorsal fin spotted.
b. Body depth 2.25-2.4 in standard length.
c. 37lateralline scales
. . . P. steindachneri (Rio
Parahyba; may be the young
of P. vimboides)
P. brevis (Bahia)
cc. 41-43 lateral line scales
bb. Body depth 2.6-3.5 in standard length.
d. 6/5-6 transverse scales (35-42 lateralline scales)
. . . P. vimboides and P. hartii
(two sibling species (?)
from Parahyba)
dd. 8-9/6-8 transverse scales.
e. Snout moderate in length, upper jaw not projecting (usually more than
47 lateral line scales) . . . P.lineatus-affinis complex
(Rio de Ia Plata, Rio
Paraguay and Rio Sao
Francisco)
projecting, resembling that of genus ee. Snout elongate, upper jaw
Ichthyoelephas.
f. Usually less than 47 lateral line scales; species from northwestern
South America . . . P.reticulatus (Venezuela)-
magdalenae(Colombia,
synonym P. beani)-
complex
219
I
..r'~.
'" ~
J-'~'-
~.~~...~
~".
Unidentified Prochilodus of the ngrcans-group. Photo by Harald
Schultz.
Young fishes of the genus Semaprochilodus, probably S. taenurus.
Photos by Dr. Herbert R. Axelrod.
Semaprochilodus sp., close to S. theraponura. Photo by Dr. Herbert.
R. Axelrod.
.--
ff; 47-48 lateralline scales; species from southeastern South America
. . . P. scrofa (Rio de Janeiro,
Rio Paraguay and (?)
Argentina)
aa. Dorsal fin plain in adults.
g. Less than 60 lateralline sca1es.
h. 41-45 lateralline scales; body depth 2.5~2.7in standard length
in adults . . . P.argenteus (type of the
genus, a synonym of P.
marggravi which is a
suppressed name;
synonyms P. /acustris and
P. costatus; Rio Sao
Francisco basin, Rio
Parnahyba)
hh. About 48-50 lateralline scales; body depth about 3 in standard
length . . . P. dobulinus (Amazon)and
P. platensis (a dubious
species from Argentina,
which may be synonymous
with P.lineatUs)
gg. About 60 lateral line scales (body depth 2.45 in standard length)
. . . P. brama (Rio Tocantins)
Profile of the heads of two bocachicos (Prochilodus) and of the patalo
(/chthyoelephas longirostris) showing the lengthening of the snout of
the latter. Prochilodus maripicru seems to be intermediary in thls
respect. Origi~al drawings by the author.
')~\ )
~~~ P.rubrotaeni~
222
g ,)
~
P.maripicru
lf
,
'i
.
(
,
..,.
.
,L
""
~
~.
~ ~
Prochilodlns (called bocachicos in Peru and jaraquis, among other
names, in Brazll) migrating upwards and jumping the falls. Photo by
Harald Schultz.
I
A young Prochilodus. Photo by Or. Herbert R. Axelrod.
Prochilodus nigricans (or P. ortonianus) from Solimoes. Photo by Or.
Herbert R. Axelrod.
.-
----
Curimata cyprinoides from Rio Purus. Photo by Or. Herbert R. Axel-
rodo
Prochilodus sp. about 5 inches in length. Photo by Aaron Norman.
GENUS ICHTHYOELEPHAS
PATALOs
The two species of the genus, 1. longirostris from Colombia and
1. humeralisfrom Ecuador, are c10selyrelated to each other; they differ in
the sizeof the eye as weIl as in a humeral spot (present in 1. humeralis).Both
are relatively large species (more than 40 em when adult), and easily
recognizable thanks to their very elongate snout.
SUBFAMIL Y CURIMATINAE
Curimatas
Unlike the bocachicos, which have been only partly reviewed, the
curimatas have been the object of a rather extensive revision in 1948.
However, this work has been much discussed, due to the extreme splitting
of the group (the number of genera having almost tripled from 10, several
already superfluous, to 27, with no less than 17 new general).
The present survey wiIl take into account only a few of these genera.
Some of the others are perhaps worthy of consideration, but this would
lead us too far astray, owing to the considerable number of species
involved (perhaps 100-110).Most of the species, like most bocachicos, lack
"good" systematic characters. They are very much alike except for the'
The most common Curimata from the Gulanas and Amazon belongs to
the C. spilura-complex. Photo by Dr. Herbert R. Axelrod, Rio Aguaro.
226
--
~
I
J
p
I
,.
-L
f
~
'\'"1
.,1 '\
.,/~
,\1
,y,.' ..
- -
Curimatopsis sp., posslbly C. macrolepis. Photo by Tlmmerman.
following features: number of scales (there are nevertheless important
overlaps), form of the preventral region (and associftted changes in the
predorsal and postventral regions, rarely that of the postdorsal region) and,
in some cases, color pattern.
AlI curimatas look like bocachicos without the predorsal spine and
without a sucking disk. The adults are toothless; the very young ones,
which probably have a very different diet, have conical teeth.
For simplicity, the curimatas will be divided into 2 groups: large species
with many scales (usually 48-110 lateral scales), and more evolutionarily
divergent species, probably derived from them. The latter are smaller, with
fewer scales, and sometimes evolve like the tetras, with the appearance of
forms with .scaled caudal fins, pigmy forms with an incomplete lateral
line, etc.
GENUS CURIMATA-typical and specialized forms
LARGEeu RIMATAS
This group inc1udes mostly forms living in large rivers, in which the
structure of their ventral region has become adapted to their way of life (a
flat breast often meaning a life near the bottom, a keeled breast an
adaptation to pelagic life, and so on). From the least specialized to the most
specialized forms (apparently), we have one large group of species with not
227
Curimata rhomboides. Photo by Dr. Herbert R. Axelrod.
~~ ---
-
,
-ri
".-
Curimataisognatha, juvenile. Photo by Dr. Herbert R. Axelrod.
Curimata (Gasterostomus) latior. Photo by Dr. Herbert R. Axelrod.
I
11
",}"*'"'
~.
,~~,
y~'
?"
.w
:":'\0
229
very many scales (48-70 lateral line scales), a short ana1 fin as in most
curimatids, and the predorsal line usually not modified; and a small group
of 4 species, placed in 3 different genera ~or subgenera, depending on the
authors), with many scales (85-110 laterI line scales), 15-17 anal rays
(except in C. p/anirostris), and the predorsalline naked.
The first sub-group inc1udes species c10selyallied to the type of the genus
Curimata, C. cyprinoides, and several related forms, as follows:
a. Postventral keel, if any, not spinous.
b. Ratber elongate species, body depth usually more tban 2.6 in standard lengtb.
c. 50-54 strongly ctenoid lateralline scales (probably no postventral keel)
. . . C. rutiloides(Mato Grosso
and Rio Negro) and C.
aspera (Rio Huallaga)
cc. 56-70 feebly crenulate lateralline scales . . . C.vittata(MatoGrosso
and Rio Negro) and C.
macrops(SaoPaulo)
bb. Rather deep species, body depth usually less than 2.6 in standard lengtb.
d. Predorsal region trenchant; lateralline relatively low, witb only 6 or 7
transverse scales to ventral fins . . . C.simulata (Amazon),
C. cisandina and C. alleni
(sympatric species from
Peruvian Amazon)
dd. Predorsal region rounded; lateral line along axis of body, witb 8-11 .
transverse scales to ventral fins.
e. Preventral region more or less rounded; scales strongly ctenoid or ciliate
(50-61Iateralline scales) . . . C.ciliata(synonyms
C. essequibensis(?), Guyana,
and C. isognatha, Amazon,
Sao Paulo ?)
ee. Preventral region fIat, the angles well marked; scales never strongly
ctenoid . . . C.cyprinoides-group,
comprising the nominal
form C. cyprinoides (from
lower Amazon), C.
cyprinoides schomburgki
(Guianas), C. copei
(Surinam), C. kneri (Rio
Negro) and C.lalcata
(Rio Xingu)
aa. Postveptral keel spinous . . . C. rhomboides-group
(Amazon basin); the 2 or
3 spp. (?) ofthis group are
often classified as
Psectrogaster, a subgenus of
Curimata or a distinct
genus
230
-:::? - ""' .,
~"l,"!i1In;rfff~
""
,
/' t
/
--
~)_)77
~
~/ ff~
.,,:rrr
~--
,4/
~<>
/,~?'-~
"r
~/
'....
,...
l'
J \
, "'- \ ....
d
% "'" '" ~.
~
,,"; ~ -\
,~;~~ ,>~
"
",.
'-
Regression of the teeth in a young, unldentifled species of Curimata
from Guyana. The first figure was drawn from a post-Iarva 15 mm in
standard length; the second from a post-Iarva 17 mm In standard
length (some teeth are already lost); and the third from a young indi-
vidual somewhat more than 20 mm in standard length, wlth the tooth-
less jaws characteristlc of adult curimatins. Original drawings by the
author, ali of the same scale.
'I
231
-.--
-:
~
Curimata sp., possibly C. metae. Photo by Or. Herbert R. Axelrod.
Curimata of the elegans-group. Photo by Or. Herbert R. Axelrod.
Curimata spilura. Photo by Or. Herbert R. Axelrod.
~..
~-
~ ,
,
Curimata vittata from the vicinity of Humaita, Brazil. Photo by Or.
Herbert R. Axelrod.
.---
'~
~~O"
"
... '.
,
rft.i:~'
~rr;~~
i!
WfIl
:' ',' i' .8:, '
~ ,/ .. ~"t'\ :' ,
~ ['W~
~
.I'!~~
~.iAf! " '~~
W-b I:. . ~ ~
~.
~
I
ro.
!'I ,.
--
,.,
232 233
,
(!I
,.
;IiiioIo...,."
11I
I!i
:7
,,'
,/
I/:~...
/ )P'
'')r
~
''''
...~ '.
".J ..
"
J
'<{.,
P' .. '
~
/(F: . .
~
~'j
.
"
.
',(
,
"
'.
~
"
'./.'"
.
,--,:,.-' ~
~
L.' >
;;tb:, c ~ . ') .,~"
.3", "V ";;
Regression of the teeth in another specles of Curimata, possibly C.
hel/eri. The first figure drawn from a post-Iarva 20 mm in standard
length; the second from a sllghtly larger individual (23 mm SL). OrIgi-
nal drawings by the author.
The second subgroup includes only a few specialized species, alI with.
a postventral keel; they may be identified as follows:
a. Preventral region rounded or flattened as in preceding subgroup.
b. Postventral keel not spinous (genus or subgenus Semitapicis).
c. 82-90 lateralline scales and 11-12 anal rays (total)
. . . C. (S.) planirostris
(Amazon, Rio Negro)
cc. 95-110 lateralline scales and 15-17 anal rays (total)
. . . C. (S.) laticeps(Amazon,
Venezuela)
bb. Postventral keel spinous (genus or subgenus Potamorhina)
. . . C. (P.) pristigaster
(Amazon, Rio Negro)
aa. Preventral region forming a continuous keel with postventral region (genus or
subgenus Gasterotomus) . C. (G.) latior (Amazon
and La Plata basins)
SMALL CURIMATAS-evolved forms
This is the largest group, with some 80-90 species. It may be split into
curimatas with a scaled caudal (corresponding to the tetra genus
Moenkhausia, for example), curimatas with an incomplete lateral line
(corresponding to the tetra genus Hyphessobrycon,for example), and
"ordinary" unrnodified curimatas (corresponding to the tetra genus
Astyanax, for example). The appropriate generic name for the last group
is debated. Curimata (in the broad sense) will be used here.
234
19
According to certain authors, most of the small curimatas fall into the
following divisions:
a. Preventral region not very flattened, usually rounded (not as in the rype species
C. cyprinoides,for example).
b. Postventral region rounded, predorsal region more or less angular.
c. Mouth terminal or subinferior: about 25 or more species, for which the
generic names Cyphocharax (oldest), A/lenina, Curimatorbis and Pseudo-
curimatahave been applied. The most common species are C. argentea, C.
elegans(which have a black spot at dorsal fin base), C. gilberti, C. spilura
(which have a more or less conspicuous caudal spot), C. microcephala,etc.
cc. Mouth inferior: about 10 species for which the generic name Cruxentina
has been suggested; the best known species is perhaps C. nasa.
bb. Postventral region keeled, predorsal region rounded: 2 or 3 species with
terminal mouth, which have 'been called by the generic name Bondia.
aa. Preventral region flattened as in the type species C. cyprinoides.
d. Mouth terminal or subinferior: a few species (3 or 4) sometimes called
Rivase/la.
dd. Mouth inferior: two species, C. hypostomaand C. trachysteta, for which
the generic name Steindachnerinahas been proposed.
The 20 or more species which are not included above would probably falI
within section (c) when the form of their ventral region (usually not
described) is studied. AIso, a specimen without an adipose fin, perhaps an
abnormal individual but possibly a good species, has been called by the
generic name Curimatoides.
Typeof Curimata multilineata. Original drawlng by Mr. Pablo Bravo In
1930, publlshed for the first time by permisslon of Dr. G.S. Myers.
235
Curimatalineopunctata, arare species from Rio Calima, Colombia.
Photo by Or. Herbert R. Axelrod.
Unidentified, parasitized species of Curimata from Rio Aguaro, pos-
sibly Curimatella alburna. Photo by Or. H.R. Axelrod.
236
..........-
'r
Curimatopsis macrolepis, male. Photo by Or. Herbert R. Axelrod.
Unidentified species of Curimata, close to C. microcephala. Photo by
Harald Schultz.
237
GENUS CURIMATELLA )
CURIMATINS WITH SCALED CAUDAL
Apart from the numerous scales on the caudallobes, this genus cannot be
told apart from Curimata. The species can be identified by means of the
following key:
a. More than 6 transverse sca1esabove lateral line; 43-46 lateral line scales.
b. 9/7 transverse sca1es; body depth 2.5-2.66 in standard length
. . . C. lepidura(RioS.
Francisco)
bb. 7t/6t transve~sesca1es;body depth 3 in standard length
. . . C. xinguensis(RioXingu)
aa. 5 or 6 transverse scales above lateral line; 28-42 lateralline scales.
c. 39 to 42 lateral line scales; body depth usually more than 3 in standard
length (no peduncular spot; a plumbeous longitudinal band)
. . . C. meyeri(Amazonand
Rio Negro)
cc. 30 to 38 lateralline sca1es;body depth usually less than 3 in standard length.
d. A peduncular spot.
e. About 35 lateralline sca1es;preventral region flattened; body elevated,
with an acute angle just in front of the dorsal fin
. . . C. dorsalis,possible
synonymisthesubspecies
C. alburna caudimaculata
(Amazon)
ee. 30 to 32 lateral line sca1es; preventral region rounded or not much
flattened; predorsal regionnot angular. . . C. rehni(Mato Grosso)
dd. No peduncular spot (35 to 38 lateral line sca1es; preventral region
flattened) . . . C.alburna,possible
synonyms C. immaculata
and (?) C. lineata (very
common in the Amazon
and Guianas)
GENUS CURIMATOPSIS
CURIMATINS WITH INCOMPLETE LATERAL UNE
As pointed out above, certain curimatas have followed parallel evolution-
ary lines with the tetras, and someof them havea very short lateralline. The
group is a difficult one. A few small species(which have been put into the
dubious genus Hemicurimata) may well be the young of large curimatas,
the lateral line usually not beipg complete at the posdarval stage. The
Hemicurimata species (esperanzae, maculata, saladensis, and possibly others)
are just like curimatas, with the lateralline perforating the first 6-10 scales
only.
The Curimatopsis species are more peculiar. They differ from Curimata
not only in their short lateralline, but also in their relanvely large maxillary
bone, which is free from the antorbital (whereas it is partially covered by_
238
\~
I.'
---
--
Curimatellaalburna (above) and C. meyeri (below). Photos by Dr. J.
Gery.
the antorbital in the other curimatins), as well as by a curious sexual
dimorphism: the males, when they reach maturity, have a much deeper
caudal peduncle than the females. Two species have been described, the
well-known Curimatopsis macrolepis from the Amazon and the Guianas
(with a dubious subspecies, C. macrocephalus,in the lower Amazon), and
the more recendy discovered C. evelynae (named for Mrs. H. R.
Axelrod),fromthe upper Rio Meta in Colombia.The malesof the latter
species, which are nicely colored with pastel tints, have no caudal spot (or a
very faint one), and are more elongate than the males of C. macrolepis.
SUBFAMILV ANODINAE
This curiousgroup includesonly3 speciesof pelagicfishes, strongly
resembling a hemiodid but lacking teeth, which are presently in an uncertain
systematic position (the resemblance with the hemiodids may be due to a
common origin* or, on the other hand, it may be due to convergence, the
samepelagicway of lifeendingin the sameadaptivestructures).
* This has just been suggested by T. Roberts (December, 1974), too late to be
taken into account in the present book.
239
~
Curimatopsis macro/epis temale. Photo by Or. Herbert R. Axel-
rodo
Curimatopsis sp. trom the Rio Negro, possibly C. eve/ynae. Photo by
Or. Herbert R. Axelrod. .
Two male Curimatopsis eve/ynae. Photo by Or. Herbert R. Axel-
rodo
Anodus e/ongatus trom Tete, Brazil. Photo by Or. Herbert R. Axelrod.
."
I
l
,,"",,,,,,,,,,,.c,,.&,.__
,
240 241
-I...-
.li
.."
Curimatella meyeri and C. rehni, type (after Fowler, 1950); and Eigen-
mannina melanopogon (photo by Dr. J. Gery).
Anodus e/ongatus(synonyms A. steatops and A. jordani) from the large
rivers in the Amazon basin (it is not found in the Guianas), and
Eigenmannina me/anopogon,this also restricted to the Amazon, are
elongate, fusiform fishes (the body depth more than 4 in the standard
length), with many small scales(95 and more), the gill membrane free from
the isthmus, a more or less developed adipose eyelid, and very numerous,
long and slender gill-rakers (55 to 100 on the lower arch), attesting to a
planktonic diet. The species differ in the form of the head, A. e/ongatus
having a normal head with the mouth terminal, whereas E. melanopogon
has a very large head as compared with the rest of the body (which tapers
caudally}, the top of the head is concave, and the mandible is produced,
with the mouth somewhat upturned. Both species are silvery, and Anodus
e/ongarushas a dark spot on the flanks (irregular in size and form), which
adds to its resemblance with a hemio'did, where this mark is common. A
third species, A. orinocensis, was described by Steindachner in 1888from
the Orinoco. It has not been rediscovered since.
242
,.....
Family Gasteropelecidae
Hatchetfishes
The hatchetfishes, or flying characins, are surface fishes identified at
once by their rounded, trenchant belly and their very large pectoral fins.
The keel is formed by the joined coracoid bones (part of the "shoulder" of
the fish), which are much flattened and enlarged and used to cleavethe water
surface. Very powerful muscles capable of setting the wing-like pectoral
fins in motion at a high speed (similar to that of the well known humming-
birds) are connected with these bones.
The technique of the flight is not entirely known. It is known that the
fish rises progressivelyabove the surface, in the manner of the ships called
hydrofoils. But one is not certain whether it really leaves the water. How-
ever, owing to the fluttering of their pectoral fins, the hatchetfishes can be
considered as true flying fishes, in contrast with pantodons and marine
"flying fishes," which only soar, and with several jumpers such as the
characoid Triportheus (which is nevertheless capable of covering several
meters above water). As for the purpose of the specialization, no one
knows. It has been suggested thtlt the fishes can eat small aerial insects or
may be able to escape predators. The author favors the latter hypothesis,
altogether without much argumento
As concems the other characters, the hatchetfishes have most in
common with the Characidae, chiefly with certain tetras. In addition, they
have the usual adptations to life at the surface (straight outline and
superior mouth).
Dr. Stanley Weitzman, one of the best comparative anatomists of our
time, has studied, among other characoid groups, the hatchetfishes.* He
recognizes two groups, which may be considered as subfamilies within the
family: the genus Thoracocharax,which is the most primitive, on the one
hand (subfamily Thoracocharacinae) and the genera Gasterope/ecusand
Carnegiella,on the other (subfamily Gasteropelecinae). Thoracocharaxis
distinguished at once by its impressive keel.
Technica1lythe genera may be identified as follows:
a. An adipose fino
b. 19-22 longitUdinalline scales; coracoid bones very enlarged (3 outer pre-
maxillary teeth; anal fin with iii, 31-42 rays)
. .. Thoracocharax
bb.28-37 longitUdinalline scales; coracoid bones less developed (0-1 outer pre-
maxillary teeth; anal fin with iii, 19-33 rays)
. . . Gasteropelecus
aa. No adipose fin (otherwise with many characters of Gasteropelecus)
. . . Carnegiella
* See for example: Tropical Fisl! Hobbyist, January, 1970, pages 38-64.
243
Thoracocharax stellatus. Photo by Harald Schultz.
Thoracocharax securis? Photo by Klaus Paysan.
.
244
,...--
I.
I
.
.
,1
Carnegiella myersi. Photo by Klaus Paysan.
Gasteropelecus sternicla. Photo by Dr. Herbert R. Axelrod.
...
''ioj
---.
I
I
245
Gasteropelecus maculatus, preserved speclmen about 21/2 Inches In
length. Photo by Or. Stanley H. Weltzman.
GENUS THORACOCHARAX
SILVERHATCHETFISHES
As stated above, the silver hatchetfishes belonging to the genus
TJwracocharaxcan be distinguished from Gasteropelecusby their more
prominent belly which forms almost a right angle with the dorsal outline.
Two species are known: TJwracocharaxstellatus, an ubiquitous fish found
from Rio Paraguay to Colombia aqd Venezuela, and Thoracocharaxsecuris
(also called pectorosus,a junior synonym), which is found mostly in the
middle-upper parts of the Amazon basin. T. stellatushas a black spot on the
base on the dorsal fin (as well as 2 or 3 rows of scales on the anal fin base
which must be counted with a lens), whereas T. securishas a plain dorsal
fin except at its tip (as well as 5 or 6 scale rows on the anal fin base).
GENUS GASTEROPELECUS
SILVE R HATCHETFISHES
This genus, formerly called Pterodiscus, includes 3 species which are very
similar in appearance. They can be distinguished from each other mostly
by anatomical details, as follows:
a. Maxillary not toothed; small size (one outer premaxillary tooth; six:h circum-
orbital bone present) . . . G.levis (Lower Amazon)
aa. Maxillary toothed; larger size.
b. Depth usually 1.8or less in standard length; body spotted; outer premaxillary
tooth and sixth circumorbital bone present as in G. levis
. . . G. maculatus (Venezuela,
Colombia, and Panama)
246
~
bb. Depth usually 1.85-1.95in standard length; body plain; no outer premaxillary
tooth; sixth circumorbital bone lacking . . . G. sternicIa(Guianasand
Trinidad; Mato Grosso in
Brazil; Peruvian Amazon)
The latter species is by far the commonest species, at least in the
aquarium trade. Almost nothing is known of its biology.
GENUS CARNEGIELLA
MARBLEO ANO PIGMY HATCHETFISHES
The species of this genus, characterized by the lack of an adipose fin, are
the smallest and prettiest of the hatchetfishes. They may be identified rather
easily by their color pattern, as well as by some technical details, as follows:
a.Body marbled with broad bands extending from keel to dorsum
. . . C. strigata (Guianas and
Amazon Basin)
aa. Body with small spots or lines.
b. Keel with narrow parallellines, sometimes with a black border; pectoral fins
with a blak band; body rather deep; circumorbital bones lacking; anal fin
with less than 31 rays; maxilla with less than 3 teeth
. . . C. marthae (Venezuela;
Uppa Amazon basin; Rio
Madeira)
bb. Keel with a black border, plain or with some dots; pectoral fins plain; body
relatively elongate (as concerns a hatchetfish); 2nd to 5th circumorbital bones
present; 33-36 anal rays; maxilla with 4-5 teeth
. . . C. myersi (Peruvian and
Bolivian~azon)
In contrast to the other species of hatchetfishes, the species of
Carnegiellaare what zoologists call polytypic, which means that 2 or more
distinct populations can be recognized within the species concepts of
Carnegiellamarthae, C. myersi and C. strigata, respectively. As concerns
Carnegiellamarthae, sometimes called the black-winged hatchetfish, there
Carnegiella myersi, live Individual about 11/4Inches In length. Photo
by Or. Stanley H. Weitzman.
..~-
'"
Carnegiel/a strigata strigata, the form called vesca by aquarists. Photo
by Or. Herbert R. Axelrod. '
Carnegiel/a strigata fasciata. Photo by Harald Schultz.
248
r
I
Tl
I
--- , ---J
Carnegiel/a marthae, form with thoracic lines. Photo by Or. Herbert R.
Axelrod.
Carnegiel/a marthae, form with thoracic spots. Photo by Or. Herbert R.
Axelrod.
249
,
is certain1ya difference in the number of anal rays between specimens from
the Rio Orinoco and Rio Negro, and speciml:;nsfrom the Peruvian Amazon
and Rio Madeira: usually 22-25 for the former and 26-30 for the latter
(total counts; based on 17 specimens from middle-upper Rio Negro, and 18
specimens from Igarape Preto on the Amazon at the Brazilian-Peruvian
border, and Guayaramirim on the Rio Madeira). The typical form from the
Orinoco-Negro Rivers has to be called Carnegiellamarthaemarthap,and the
form from the Peruvian Amazon, which has already been described, may be
called Carnegiellamarthae schereri.
The pigmy hatchetfish, Carnegiellamyersi, seems to be polytypic too. In
Boliviathere is a population which 'is c1earlydeeper bodied than the typical
form from the Peruvian Amazon.
Finally, the present author has recently shown that the well known
polymorphism of the marbled hatchetfish Carnegiellastrigata (which has
different patterns of the thoracic bands depending on the localities) may be
explained by the existence of two forms (probably subspecies) which
perhaps intergrade, at least in some places in the Peruvian Amazon. Thanks
to infrared photography of the types in the British Museum, it has been put
into evidence that these types have the color pattern called vesca by most .
ichthyologists and aquarists. Therefore the nane C. vescais a synonym of
the older name C. strigata, and the fishes having the vesca-pattern are
rightly known as Carnegiella strigata. The fishes with the "old" strigata
pattern would be given the next available name, C. fasciata or more
precisely Carnegiellastrigatafasciata, since it is probably a subspecies and
not a "fuII" species.
In the nominate subspecies, C. strigata strigata, the foremost band,
which runs obliquely from the middle of the thoracic convexity toward a
point behind the pectoral root, is divided into two branches almost from its
originoThis pattern is seen in fishes coming from places situated north of
the Amazon, in the Guianas, upper Amazon, and also in the Rio Puros. In
the subspecies C. strigatafasciata, the thoracic band is not divided until it
reaches the leveI of the pectoral finoThis form occurs in the Amazon basin
(lower, middle and upper Amazon) and somewhat south of it, as well as in
the Rio Caqueta in Colombia, north of the Amazon. -
Both forms have been collected in the middle Amazon, in the upper Ama-
zon and in the Rio Puros. It is not known whether they live together or
isolated from each other; some samples* are intermediate in certain
characters and may prove to be intergrades. Finally, within each sub-
species there is a variability in the color pattern as well as in the body
proportions and in the anal ray COunts.
* In theRioMeta,atributaryoftheRioOrinocoinColombia, aswellasintheRio
Negro and (?) Upper Amazon. .
250
1