PROC. BIOL. SOC. WASH. 103(3), 1990, pp. 525-542 MOOJENICHTHYS MIRANDA-RIBEIRO (PISCES: OSTARIOPHYSI: CHARACIDAE), A PHYLOGENETIC REAPPRAISAL AND REDESCRIPTION Ricardo M. C. Castro and Richard P. Van Abstract.— The Neotropical characiform characid genus Moojenichthys Mi- randa-Ribeiro is hypothesized to form a monophyletic lineage with Triportheus Cope on the basis of shared derived features of the pectoral girdle and perhaps of the first infraorbital. Autapomorphies for Moojenichthys are discussed. Moo- jenichthys myersi Miranda-Ribeiro, the only member of the genus, is rede- scribed. This species is apparently endemic to the Rio do Bra^o system of the state of Bahia, Brazil. Resumo. — Uma hipotese de relafoes filogeneticas reunindo em um so grupo monofiletico os generos neotropicais Moojenichthys e Triportheus, ambos per- tencentes a familia Characidae e a ordem Characiformes, e formulada com base na posse em comum de caracteres derivados na cintura escapular e, pos- sivelmente, primeiro infra-orbital. Moojenichthys myersi Miranda-Ribeiro, o unico membro do genero, e redescrito. Esta especie e, aparentemente, endemica da bacia do Rio do Brago, no estado da Bahia, Brasil, e consiste na unica ocorrencia de um representante da linhagem evolutiva Triportheus-Moojenich- thys nos rios costeiros do leste do Brasil, excluindo o rio Sao Francisco. Miranda-Ribeiro (1956:546) proposed the was "related" to Triportheus Cope (1872) characid genus Moojenichthys for a single and Clupeacharax Pearson (1924) both of species, M. myersi, first described in that which also have elongate, laterally com- publication on the basis of two specimens pressed bodies characterized by varyingly collected in the Rio do Bra90, near Ilheus developed mid-ventral keels. Miranda-Ri- in the state of Bahia of eastern Brazil. Af CO- beiro did not explicitly state which char- jenichthys myersi has a number of external acters lead him to propose that these species anatomical features unusual within the are "related," presumably closely. It seems Characiformes. Perhaps the most striking reasonable, nonetheless, to assume that the of these are its elongate, laterally flattened overall similarities in body shapes, partic- body, and the pronounced, laterally com- ularly the presence of mid-ventral keels, pressed mid- ventral keel. Although the tho- contributed significantly to this hypothesis, racic keel is not obvious in specimens of Despite the distinctive external mor- Moojenichthys under 15 mm SL, mid- to phology of Moojenichthys myersi, subse- large-sized individuals of the genus have a quent references to the genus and species distinct mid- ventral ridge extending from are extremely limited. Gery (1972:55), in the isthmus posteriorly to the origin of the his key to New World characiforms, com- pelvic fin. mented that Triportheus and Moojenichthys In his original description of Moojenich- are probably derived from Brycon Miiller thys Miranda-Ribeiro stated that the genus and Troschel, and on the next page pro- 526 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON posed that Clupeacharax is close to Tripor- theus (1912:56). More recently, Gery (1977: 346) again emphasized the similarities be- tween Moojenichthys and Triportheus, not- ing, however, that the reported dentition of M. myersi is more reminiscent of the subfamily "Tetragonopterinae" of the fam- ily Characidae. In that publication Gery did not comment on Miranda-Ribeiro's pro- posal of a possible relationship between Moojenichthys and Clupeacharax, but rath- er segregates Clupeacharax in the monotyp- ic subfamily Clupeacharacinae. Castro (1981:138), in turn, cited some external similarities between Clupeacharax and En- graulisoma Castro. Other than for the original description by Miranda-Ribeiro (1956) and Gery's brief comments (1972, 1977), we know of no published citations of Moojenichthys. Sim- ilarly, the primary ichthyological literature apparently does not include records of the subsequent capture of the species. This is not surprising given that M. myersi is ap- parently endemic to the Rio do Bra^o sys- tem, a poorly sampled river basin that drains into the Atlantic Ocean slightly north of the city of Ilheus in the state of Bahia, Brazil. Recent collecting efforts in the coastal rivers of Bahia associated with our revi- sionary studies of the characiform families Curimatidae and Prochilodontidae, have resulted in the capture of a large series of Moojenichthys myersi with a much greater range of standard lengths than available to Miranda-Ribeiro. This additional material allows us to provide a detailed redescription of the genus and species. Those specimens also permit anatomical studies to evaluate previous suggestions about the relation- ships of Moojenichthys. Methods.— AW measurements are given as proportions of standard length (SL) ex- cept for subunits of the head which are pre- sented as proportions of head length (HL). Lateral-line scale counts include all pored scales along that series, including the scales posterior of the hypural joint. Vertebral counts were taken from radiographs, and specimens cleared and counterstained for bone and cartilage. The vertebral count in- cludes the four vertebrae incorporated in the Weberian apparatus, and considers the fused PUi +Ui as a single element. In counts of median and pelvic fins, lower-case Ro- man numerals indicate unbranched rays, and Arabic numerals indicate branched rays. The range for each meristic value of all mea- sured specimens is presented first, with the value for the holotype indicated in square brackets. Measurements were made follow- ing the methods outlined in Fink & Weitz- man (1974:1-2). The following institutional abbreviations are used: ANSP— Academy of Natural Sci- ences of Philadelphia; MNRJ— Museu Na- cional, Rio de Janeiro; MZUSP— Museu de Zoologia, Universidad de Sao Paulo; USNM— National Museum of Natural His- tory, Smithsonian Institution, Washington, D.C.; and FFCLRP-USP-Faculdade de Filosofia, Ciencias e Letras de Ribeirao Pre- to, Universidade de Sao Paulo. Phylogenetic Analysis As noted in the introductory comments, Miranda-Ribeiro (1956:546) considered Moojenichthys, Triportheus, and Clupea- charax to be related, albeit without specif- ically stating the basis for his opinion. A number of questions exist about the phy- logenetic associations of these taxa, and at least in the case of Triportheus the species- level classification of the genus remains un- settled. Moojenichthys myersi has been collected only in the Rio do Brago system of eastern Brazil. Clupeacharax includes a single species, C. anchoveoides, a poorly known fish reported from scattered sites ranging from Argentina (Miquelarena & Casciotta 1982:333), through Bolivia (Pearson 1924: 47) and Peru (Ortega & Vari 1986:8), to Ecuador (Stewart et al. 1987:26). Triporthe- us, a much more speciose genus well rep- VOLUME 103, NUMBER 3 527 resented in museum collections, is found on both sides of the Andean Cordilleras across much of lowland South America. Miranda- Ribeiro (1941), who applied Chalcinus Cu- vier & Valenciennes (1 849) to the members of Triportheus, recognized eleven species in- cluding Chalcinus culter Cope which Fowler (1907) segregated in the genus Coscinoxy- ron. Myers (1940:170) pointed out that Chalcinus was already occupied in the Hy- menoptera, and that Triportheus was the next available name for the members of the genus. Schultz (1944:273) and Weitzman (1960:239) agreed that Miranda-Ribeiro's revisionary effort was unsatisfactory, a view that we share. Gery (1977:343, 654) in his key to the members of Triportheus, recog- nized only nine of the nominal species as valid, and tentatively retained culter in Triportheus. This uncertainty concerning the recognizable species of Triportheus compli- cates a determination of whether the genus is monophyletic. Those questions are be- yond the scope of this study. We will, rather, attempt to determine the phyletic relation- ships of Moojenichthys, and critically eval- uate suggestions that the genus is related to Triportheus and Clupeacharax. Mid-ventral keel.— One of the most ob- vious features of Moojenichthys is the dis- tinct mid-ventral keel that extends poste- riorly from the isthmus to between the origins of the pelvic fins. Whereas the pos- terior portion of this keel is formed by fleshy tissue, the anterior portion is underlain by asymmetrically expanded coracoid bones. The somewhat irregular anterior margins of the coracoids are relatively short where they meet the anteroventral margins of the cleithra (Fig. IB). The dorsal margin of the coracoids in Moojenichthys gradually rises posteriorly to the region where it articulates with the scapula and mesocoracoid. As a consequence, the overall proportions of the plate-like ventral portion of the coracoid are distinctly asymmetrical in lateral view, with the posterior margin of the coracoid notably deeper than the anterior border of the bone. Such asymmetrically expanded coracoids are relatively unusual within characiforms. A horizontally rectangular, moderately-sized coracoid with an overall horizontally rect- angular form is found in the vast majority of characiforms of all families (e.g., Charac- idae (Brycon, Fig. lA); Prochilodontidae, see Roberts 1973:fig. 24; Hemiodontidae, see Roberts 1974:figs. 16, 53; Lebiasinidae, see Weitzman 1964:fig 10; Parodontidae, see Roberts 1974:fig. 76; and Erythrinidae, see Starks 1930:fig. 8). Indeed, only the cha- raciform families Characidae and Gaster- opelecidae include taxa in which the cora- coids are dramatically expanded to form distinct thoracic keels. Within the Charac- idae enlarged coracoids are absent among Old World members of the family, and the vast majority of Neotropical characids sim- ilarly lack significant expansions of these bones (e.g., Acestrorhynchus, see Roberts 1969:fig. 52; and Brycon (Fig. lA), see also Weitzman 1962:figs. 18, 19). As noted in the introductory section, the species of Trip- ortheus are also characterized by a distinct thoracic keel. The coracoids in Triportheus are similar to those of Moojenichthys in being asymmetrically and vertically ex- panded in lateral view. In Triportheus, how- ever, the dorsal margin of the plate-like ven- tral portion of the ossification is more steeply angled, and the posterior margin of the bone much more extensive vertically than that in Moojenichthys (Fig. IC). This gives the pro- file of the bone a near equilateral triangular appearance. Given that most characids and non-characid characiforms lack enlarged coracoids, the common occurrence of ex- panded coracoids found in Moojenichthys and Triportheus is reasonably hypothesized as a derived character which, in turn, would be congruent with the hypothesis that the two genera are sister taxa. Expanded coracoids associated with a thoracic keel are, however, not unique to Moojenichthys and Triportheus among characiforms. As noted in the introductory discussion, an enlargement of that element 528 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON B C Fig. 1. Coracoids of A) Brycon falcatus, USNM 226161, 75.1 mm SL; B) Moojenichthys myersi, USNM 304497, 84.8 mm SL; and C) Triportheus angulatus, USNM 270343, 76.6 mm SL; right side, medial view, anterior to left. also occurs in Clupeacharax and may have lead Miranda-Ribeiro to suggest that the ge- nus was related to Moojenichthys. Starks (1930:22-23, fig. 9) described and figured the expanded coracoid of Rhaphiodon Agassiz, a genus of large predatory Neo- tropical characiforms. Weitzman (1960:239) noted that a keeled thorax and expanded coracoids, although unusual among characi- forms (his "characids"), also occur in Pia- bucus Oken, Pseudocorynopoma Perugia, Rhaphiodon, and the genera Carnegiella Ei- genmann, Gasteropelecus Pallas, and Thor- acocharax Fowler, the last three of which together constitute the Gasteropelecidae (sensu Greenwood et al. 1966:395). Ex- panded coracoids also are found in Cynodon Spix, Hydro lycus Miiller and Troschel and Gnathocharax Fowler. All of the above cit- ed taxa appear to be surface feeders which. with the exception of the Gasteropelecidae, also have moderately to distinctly elongate bodies. The occurrence of expanded coracoids in a number of characid taxa other than Moo- jenichthys and Triportheus brings, into question the appropriateness of using that feature to propose a close relationship for those genera. Weitzman suggested (1954: 230-23 1) that the specialized expansions of the coracoids probably arose several times within the Characiformes. The present poor understanding of phylogenetic relationships within the Characiformes in general, and the Characidae in particular, restricts the degree to which we are able to critically evaluate that suggestion in all instances. Nonetheless, subsequent research by var- ious researchers has yielded data that sup- port Weitzman's suggestion, and which in- VOLUME 103, NUMBER 3 529 dicates that the occurrence of expanded coracoids in characids other than Moojen- ichthys and Triportheus is homoplastic rel- ative to that feature in those genera. Vari (1977:4-6) discussed a series of dis- tinctive derived features of the posterior chamber of the gas-bladder and of the an- terior proximal pterygiophores of the anal fin that unite Piabucus with Iguanodectes Cuvier. In Iguanodectes the pre-pelvic re- gion is transversely rounded. Pseudocory- nopoma is a member of the subfamily Glan- dulocaudinae, a taxon which Weitzman et al. (1985:1 12-1 1 3) noted may not represent a monophyletic assemblage. Nonetheless those authors tentatively suggested that Pseudocorynopoma may be part of a mono- phyletic subset of genera within the Glan- dulocaudinae. Other species in the subfam- ily lack keeled thoracic regions. Howes (1976) united Rhaphiodon, Cynodon, and Hydrolycus as a tribe, the Cynodontini, and hypothesized on the basis of a variety of characters that the lineage consisting of these three genera was most closely related to var- ious genera of the characid tribe Characini. The Characini, in turn, consists of species in which the coracoids are not dramatically enlarged. Thus the available evidence indicates that the characid taxa with keeled thoracic re- gions cited in the previous paragraph are each in turn most closely related to species or species groups without that derived mod- ification. The species with expanded cora- coids cited in the immediately preceding section also lack the derived features of the infraorbital series and lateral ethmoid com- mon to Moojenichthys and Triportheus (see discussions in following sections). Conse- quently it is most parsimonious to assume that the pre-pelvic keels of Piabucus, Pseu- docorynopoma, Rhaphiodon, Cynodon, and Hydrolycus are homoplastic relative to the expanded coracoids in Moojenichthys and Triportheus. Information concerning the phylogenetic relationships of the Gasteropelecidac, Clu- peacharax, and Gnathocharax is somewhat more equivocal. The three genera of the Gasteropelecidac constitute a highly de- rived lineage presumably derived from some component of what is now recognized as the Characidae (Weitzman 1954:243). Al- though the closest relatives of the Gaster- opelecidac remain to be elucidated, we agree with Weitzman (1954) that the overall char- acters of gasteropelecids differ dramatically from those of Triportheus and that gaster- opelecids are apparently evolved from a dif- ferent subunit within the Characiformes. Although a resolution of the phylogenetic associations of gasteropelecids would re- quire an analysis that extends far beyond the scope of this paper, one feature of the Gasteropelecidac is noteworthy relative to this question. Weitzman & Fink (1983:39 1) noted that the supraorbital bone is absent in all the "tetragonopterine" characids they examined. The supraorbital is widely dis- tributed among characiforms, and also among the members of the Characidae both in the New World and Africa. The absence of that ossification is thus hypothesized to be a derived condition within the Charac- idae. The supraorbital is absent in the Gas- teropelecidac (see Weitzman 1954:7), and the common absence of the supraorbital in "tetragonopterines" and gasteropelecids may be a derived feature indicative of com- mon ancestry of those taxa. Both Triporthe- us and Moojenichthys, in contrast, retain a supraorbital. The relationships of Gnathocharax, a monotypic genus of the Amazon basin with an expanded coracoid, are still unresolved. Gnathocharax lacks the distinct anterior process of the lateral ethmoid common to Moojenichthys, Triportheus, Brycon, and various other characids (see discussion un- der "Autapomorphies of Moojenichthys""). Gnathocharax also does not have the dis- tinctive modification of the first infraorbital found in Moojenichthys, Triportheus, and at least some species of Brycon. Finally, Gnathocharax lacks a supraorbital and is 530 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON characterized by conical dentition typical of the characid tribe Characini and groups probably aligned with that tribe (see also discussions concerning the monophyly of the Characini by Menezes (in Sazima 1983), Vari (1986), and Weitzman & Vari (1987)). The cumulative data is congruent with the hypothesis that the coracoid expansion in Gnathocharax is homoplastic relative to that in Moojenichthys and Triportheus. The phylogenetic relationships of Clu- peacharax, the last characid genus in our list of genera with expanded coracoids are uncertain. Castro (1981:138) noted a series of similarities between Clupeacharax and Engraulisoma which has non-expanded coracoids. Ongoing studies by one of us (RMCC) are aimed at analyzing the signif- icance of these similarities between Clupea- charax and Engraulisoma. In the interim nonetheless, several features of Clupea- charax bring into question the hypothesis of a close relationship between that genus and Moojenichthys as first proposed by Mi- randa-Ribeiro (1956). Clupeacharax lacks the derived form of the first infraorbital common to Moojenichthys, Triportheus and some species of Brycon (see discussion un- der "Infraorbitals"). Furthermore, Clupea- charax lacks the anterior process of the lat- eral ethmoid common to Triportheus, Moojenichthys, Brycon, and various other characids (see discussion under "Autapo- morphies of Moojenichthys'''). Thus even in the absence of a detailed analysis of the re- lationships of Clupeacharax it is more par- simonious to hypothesize that the enlarge- ment of the coracoids in that genus are homoplastic with respect to those in Moo- jenichthys and Triportheus. In summary, the presence of expanded coracoids in gasteropelecids and various characids besides Moojenichthys and Trip- ortheus, thus appears to have arisen inde- pendently of that feature in those genera. The possession of asymmetrically vertically enlarged coracoids is consequently pro- posed as a synapomorphy for Moojenich- thys and Triportheus. In our introductory discussion we note that Gery (1977:346) commented that the dentition of Moojenichthys is "approaching [that of] the Tetragonopterinae." The pres- ence of a supraorbital in Moojenichthys myersi argues, however, against a close phy- logenetic alignment of Moojenichthys with tetragonopterines in which that ossification is apparently absent (Weitzman & Fink 1983:391). Infraorbitals.— K second possible syn- apomorphy for Triportheus and Moojenich- thys is found in the infraorbital series. Moo- jenichthys and Triportheus have the series of ossifications surrounding the orbit typical for characiforms, six infraorbitals, a su- praorbital, and an antorbital. The most no- table feature in these series of ossifications in Moojenichthys and Triportheus involves the form of the first infraorbital and its as- sociation with the second infraorbital. The posteroventral portion of the first infraor- bital in both Moojenichthys and Triportheus extends distinctly ventral of the antero ven- tral portion of the second infraorbital there- by significantly reducing the degree to which the latter element enters into the outer mar- gin of the infraorbital series (Fig. 2). This association of the two anteriormost infraor- bital elements differs from the morphology of these bones in Clupeacharax, Rhaphio- don, Piabucus, Pseudocorynopoma, Carne- giella, Gasteropelecus, Thoracocharax, Cynodon, Hydrolycus, Gnathocharax, the other characid genera known to have ex- panded coracoids. In those other taxa the first and second infraorbitals meet along a straight anteroventrally to posteroventrally aligned juncture without any invasion by the first infraorbital of the area primitively occupied by the antero ventral portion of the second infraorbital. Although the derived form of the first and second infraorbitals in Moojenichthys and Triportheus serves to distinguish those taxa VOLUME 103, NUMBER 3 531 ANT Fig. 2. Infraorbitals and supraorbital of Moojenichthys myersi, USNM 304497, 84.8 mm SL, right side, anterior to right. Abbreviations: ANT— antorbital; lO— infraorbitals (1 to 6); SO— supraorbital. from other characids with expanded cora- coids, we should note that the occurrence of such modifications of the infraorbitals extends beyond Moojenichthys and Tripor- theus. Howes (1982:5) illustrated a mor- phology of the first and second infraorbitals in Brycon acuminatus (Eigenmann & Nor- ris) apparently comparable to that in Moo- jenichthys and Triportheus. At least some other species of Brycon, in contrast, have more generalized associations between the first two infraorbitals (e.g., B. meeki, see Weitzman 1962:fig. 8). Brycon has been suggested to be a likely close relative to Triportheus by both Regan (191 1:18) and Weitzman (1960:243). In his overview of the former genus Howes (1 982: 1) questioned whether Brycon is monophy- letic, but utilized the traditional concept of the genus until future phylogenetic and re- visionary studies can be undertaken. In the absence of such phylogenetic data Howes excluded Triportheus from Brycon on the basis of the presence of the mid-ventral keel in the latter genus; noting that Triportheus shares, however, all the other diagnostic characters of Brycon. The various similar- ities between Triportheus and Brycon noted 532 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON by Regan, Weitzman, and Howes, and the derived features of the infraorbitals cited above for Moojenichthys, Triportheus, and at least some species of Brycon, raise the question of whether the lineage formed by Triportheus and Moojenichthys may be most closely related to some subunit of Brycon. Resolution of that question must await fur- ther revisionary and phylogenetic studies of both Triportheus and Brycon. Autapomorphies of Moojenichthys my- ersi.— Within the lineage formed by Moo- jenichthys, Triportheus, and possibly Bry- con or some subunit of the latter genus (see immediately preceding discussion), one of the more notable derived features for Moo- jenichthys myersi (hereafter referred to as Moojenichthys) is the absence of the single symphysial tooth posterior to the main row of dentition on each dentary. Such sym- physial teeth are common to all members of Triportheus and Brycon, and also occur in Chalceus Cuvier. Chalceus shares all the defining characters of Brycon, but is ex- cluded from Brycon because it possesses a supramaxilla (Howes 1982:1-2). An inner row of teeth on the dentary developed to varying degrees also occur in various char- aciforms in both the New World (e.g., Le- biasinidae, Weitzman 1 964: 1 43) and Africa (e.g., diverse genera in the Characidae, see Poll 1957:95, and Distichodontidae, Vari 1979:275-277). This broad phyletic distri- bution of symphysial teeth both in groups proximate to Moojenichthys and other more distantly related characiforms, makes it most parsimonious to hypothesize that the ab- sence of the symphysial dentary teeth in Moojenichthys is a derived loss. The ventrally recurved form of the max- illary dentition and the large number of teeth along the anterior margin of the maxilla dis- tinguish Moojenichthys (Fig. 3) from all examined species of Triportheus. The ven- trally recurved maxillary teeth in Moojen- ichthys are unique within the assemblage formed by that genus, Triportheus, and pos- sibly Brycon, and are hypothesized to be an autapomorphy for Moojenichthys. In con- trast, the large number of teeth ( 1 4 to 20) along the anterior margin of the maxilla in Moojenichthys is more difficult to evaluate. On the one hand that dentition does serve to readily separate Moojenichthys from Triportheus which has only 2 to 4 teeth in that series. Alternatively the species of Bry- con have 10 to 30 teeth along the maxilla (Howes 1982:46), with most species over- lapping the range in tooth number oi Moo- jenichthys to some degree. This common occurrence of large numbers of maxillary teeth in Moojenichthys and Brycon raises the possibility that the relatively few max- illary teeth in Triportheus may be synapo- morphic for the members of that genus, and that the high number of maxillary teeth in Moojenichthys is primitive. A final noteworthy autapomorphy for Moojenichthys involves the anterior portion of the lateral ethmoids. In Moojenichthys, Brycon and Triportheus the anterior surface of the lateral ethmoid bears a distinct pro- cess that extends anteriorly and medially to contact the posterodorsal surface of the vomer (see Weitzman 1962:fig. 3 for an il- lustration of the condition in Brycon meeki). In Moojenichthys the anterior process of the lateral ethmoid is developed into an elon- gate anteriorly-tapering process (Fig. 4) that is significantly longer than comparable pro- cesses in Triportheus and Brycon. Anterior- ly this elongate anterior process of the lat- eral ethmoid in Moojenichthys contacts a distinct lateral process situated on the dor- sal surface of the vomer. The space between the anterior processes of the paired lateral ethmoids and dorsal of the vomer and para- sphenoid is, in turn, filled by a large cartilage mass comparable to that in Brycon, Tripor- theus, and many other characids. Although the presence of the anterior process of the lateral ethmoid is not unique to Moojenich- thys, the degree of the anterior elongation of the structure is not equalled in other ex- amined characiforms and this modification is thus considered autapomorphic for the genus. The relationship of the anterior process VOLUME 103, NUMBER 3 533 PMX DEN Fig. 3. Upper and lower jaws of Moojenichthys myersi, USNM 304497, 84.8 mm SL; left side, anterior to left; individual bones separated from positions in life. Abbreviations: DEN— dentary; MX — maxilla; PMX— premaxilla. of the lateral ethmoid to the parasphenoid, vomer, and associated median cartilage in Moojenichthys is also distinctive. When present, the anterior process of the lateral ethmoid in characids usually extends di- rectly along the lateral margin of the para- sphenoid and vomer (e.g., Brycon meeki, see Weitzman 1962:fig. 3). Moojenichthys, in contrast, has a distinct vertical gap be- tween the ventral margin of the anterior process of the lateral ethmoid and the dorsal surface of the vomer. This results in a broad lateral exposure of the median cartilage mass (Fig. 4). This relationship of the lateral eth- moid and proximate bones and cartilages is unknown in Triportheus, Brycon and other examined characiforms, and is consequent- ly hypothesized to represent an additional autapomorphy for Moojenichthys. Moojenichthys Miranda-Ribeiro Moojenichthys Miranda-Ribeiro 1956:546 [type Moojenichthys myersi Miranda-Ri- beiro, by original designation]. — Gery 1972:55 [possible derivation from Bry- con].— Gcry 1977:346 [similarities with Triportheus noted; dentition compared with that of "Tetragonopterinae"]. 534 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON LE APLE CART PARA Fig. 4. Anterior portion of the neurocranium of Moojenichthys myersi, USNM 304497, 84.8 mm SL; right side, lateral view, anterior to right. Abbreviations: APLE— anterior process of lateral ethmoid; CART— cartilage; LE— main body of lateral ethmoid; PARA— parasphenoid; RH— rhinosphenoid; VO— vomer. Diagnosis.— 'Within the clade formed by Moojenichthys and Triportheus only the for- mer genus lacks the paired symphysial teeth posterior to the main row of dentary den- tition. Moojenichthys also differs from Trip- ortheus in the numerous ventrally recurved teeth along the anterior margin of the max- illa, and in the derived degree of develop- ment and position of the anterior process of the lateral ethmoid. Moojenichthys can also be distinguished from Triportheus on the basis of its possession of two rather than three rows of teeth on the premaxilla, and in having numerous teeth along much of the anterior margin of the maxilla rather than several teeth limited to the dorsal portions of the maxilla. Remarks. —The derived features de- scribed above are congruent with the hy- pothesis that Moojenichthys forms a mono- phyletic lineage with Triportheus, or with Triportheus and a subunit of Brycon. This conclusion raises the question of whether it is appropriate to continue to recognize a monotypic Moojenichthys, or whether that genus should be synonymized into Tripor- theus. Moojenichthys is characterized by a series of autapomorphic features. Within the Characiformes such phenetically distinct VOLUME 103, NUMBER 3 535 taxa have been traditionally segregated into monotypic genera. We reject the criterion of phenetic distinctness as an a priori basis for the continued recognition of Moojenich- thys, but nonetheless suggest that it is pre- mature to synonymize Moojenichthys into Triportheus. Our decision is a consequence of the present poor knowledge of the phy- logenetic relationships within Triportheus. Until such time as we have a rigorous hy- pothesis of the relationships within Tripor- theus it is impossible to hypothesize wheth- er Moojenichthys is the sister-group to that genus, and thus could be preserved, or sim- ply represents a subunit of Triportheus. In the latter case the continued recognition of Moojenichthys would result in a paraphy- letic Triportheus. That problem could be re- solved in one of several ways depending on the topology of the phylogenetic tree for the clade consisting of Moojenichthys plus Trip- ortheus. Moojenichthys could be synony- mized into Triportheus along with Coscino- xyron, a genus not presently recognized by most authors. Alternatively the topology of the phylogenetic tree might be such that it would be possible to continue to recognize Moojenichthys either by expanding the def- inition of the genus, or by recognizing one or more other genera in the clade in addition to Triportheus and perhaps Coscinoxyron. Given these diverse possibilities, we con- tinue to use Moojenichthys in this study. Moojenichthys myersi Miranda-Ribeiro Figs. 1-7, Table 1 Moojenichthys myersi Miranda-Ribeiro, 1956:546-547, fig., type locality: "Bra^o River, Ilheos [=Ilheus], state of Bahia, Brazil.— Gery, 1977:346 [citation; pos- sible relationships]. Diagnosis.— See "Diagnosis" of Moojen- ichthys above. Description. —Morphometries of holo- type, paratype and larger examined non-type specimens presented in Table 1 . Body elon- gate, distinctly compressed laterally in all specimens greater than 25 mm SL, some- what less so in smaller individuals. Greatest body depth located slightly anterior to ver- tical line through origin of pelvic fin, ap- proximately equal to one-half length of longest pelvic-fin ray in specimens over 50 mm SL; body not as deep and mid-ventral keel less developed in smaller examined specimens. Dorsal profile of head slightly convex from margin of lip to vertical line through posterior nostril, nearly straight from that line to rear of head. Dorsal profile of body slightly convex from rear of head to origin of dorsal fin, posteroventrally slanted and somewhat convex along base of dorsal fin; straight from posterior termina- tion of dorsal fin to adipose fin, and mod- erately concave along caudal peduncle. Dor- sal portion of body obtusely keeled transversely anterior to dorsal fin; trans- versely rounded posterior to fin. Ventral profile of head distinctly convex over lip, straight along antero ventral margin of jaw, and distinctly convex ventral to joint with quadrate. Ventral profile of body irregular, distinctly convex overall; very slightly con- vex from isthmus nearly to vertical line through origin of pectoral fin; convexity greater from that line to origin of pelvic fin; straight to slightly concave from origin of pelvic fin to anterior termination of anal fin; straight and posterodorsally slanted along base of anal fin; slightly concave along dor- sal peduncle. Distinct mid-ventral keel ex- tending from isthmus to between origins of pelvic fins; keel less developed in specimens under 20 mm SL; increasingly obvious in individuals over 30 mm SL, most devel- oped in specimens of over 50 mm SL. Scales along margin of keel flat, not folded over edge of keel. Head obtusely pointed in profile; mouth terminal, lower jaw longer than upper, with dentigerous portion of maxilla distinctly an- gled posteroventrally. Maxilla extending posteriorly under orbit to vertical line through anterior margin of pupil. Nostrils of each side close together; anterior opening 536 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON Table L— Morphometries of Moojenichthys myersi. Standard length is expressed in mm; measurements 1 to 14 are percentages of standard length; 15 to 18 are percentages of head length. Dashes indicate measurement that could not be taken due to damage to holotype. Range includes values for 24 specimens (holotype, MNRJ 4127; paratype MNRJ 4128; and 22 of the larger non-type specimens out of USNM 304497, MZUSP 40227, MNRJ 1 1605, and ANSP 164288), with the exception of length of the longest dorsal-fin ray which is based on 23 specimens, and length of the longest pectoral-fin ray which is based on 22 specimens. Holotype Paratype Range Mean Standard length 96.6 85.8 27.0-96.6 69.1 1. Greatest body depth 25.7 25.5 22.2-28.7 27.0 2. Snout to dorsal-fin origin 65.2 62.0 60.7-65.4 62.8 3. Length of base of dorsal fin 8.2 7.5 7.1-8.7 7.8 4. Posterior terminus of dorsal fin to adipose fin 18.0 17.8 16.4-19.8 18.0 5. Posterior terminus of dorsal fin to caudal-fin base 28.9 29.6 27.9-32.4 29.8 6. Snout to origin of pelvic fin 49.6 50.5 46.7-51.3 49.2 7. Snout to origin of anal fin 68.7 67.6 62.2-68.7 65.3 8. Length of base of anal fin 31.5 28.3 28.1-32.1 30.4 9. Length of caudal peduncle 8.7 9.8 8.0-10.6 9.2 10. Length of longest dorsal-fin ray 16.8 15.3 15.3-18.6 16.8 11. Length of longest pectoral-fin ray — 32.1 24.4-32.1 30.1 12. Length of longest pelvic-fin ray 14.7 13.6 11.1-15.0 13.5 13. Least depth of caudal peduncle 7.6 7.0 6.3-8.7 7.9 14. Head length 24.1 24.2 23.7-29.0 24.6 15. Snout length 23.6 22.6 20.0-27.0 23.5 16. Orbital diameter 36.1 36.5 31.7-38.9 36.3 17. Postorbital head length 38.2 36.5 29.3-41.8 38.1 18. Interorbital width 24.0 24.0 19.1-27.6 24.5 circular, posterior kidney-shaped. Eye rel- atively large, without adipose eyelid. Me- dian fronto-parietal fontanel well devel- oped; completely separating parietals; frontals in contact only anteromedially and at epiphyseal bar. Fontanel becoming pro- gressively wider posteriorly, extending onto dorsomedial surface of supraoccipital. Infraorbital series complete (Fig. 2), all infraorbitals with laterosensory canal seg- ments. Sixth infraorbital (dermosphenotic) with single tubular laterosensory canal seg- ment. First infraorbital expanded antero- ventrally, with distinctly convex anterior margin, anterior portion extends over lat- eral surface of maxilla; laterosensory canal segment with three sections in larger spec- imens. Supraorbital and antorbital present. Four branchiostegal rays, first three at- tached to anterior ceratohyal, fourth to pos- terior ceratohyal. Gill-rakers relatively elongate, 17 or 18+1+40 to 42 rakers on outermost gill-arch (in 2 larger cleared and counterstained specimens). Lower jaw with one row of 12 to 14 teeth on each side on each dentary (Fig. 3); num- ber of teeth greater in largest specimens; in- ner row consisting of single symphysial tooth absent. Anterior 5 teeth on dentary notably larger than remainder, with 5 cusps, medial cusp distinctly largest. Remaining teeth usually tricuspidate, rarely unicuspidate, with largest cusp recurved somewhat pos- teriorly. Teeth on premaxilla in two rows; teeth of inner row larger (Fig. 3). Four tri- cuspidate teeth of approximately equal size in outer row. Six teeth in inner row on pre- maxilla; 2 medial teeth largest, subequal; remaining teeth gradually becoming smaller laterally; medial tooth tricuspidate, remain- ing teeth in row with 5 cusps. Lateral tooth of inner row of premaxilla approximates dorsal tooth on maxilla. Anterior margin of maxilla distinctly convex, with single row VOLUME 103, NUMBER 3 537 Fig. 5. Moojenichthys myersi, Brazil, Bahia, "Bra90 river" [=Rio do Bra^o], Ilheos [=Ilheus]; MNRJ 4127, holotype, 96.6 mm SL. of teeth (Fig. 3). Teeth on maxilla distinctly smaller than smallest tooth on premaxilla; typically unicuspidate in specimens of ap- proximately 27 mm SL, bicuspidate or usu- ally tricuspidate in specimens of about 40 mm SL and greater. Smaller specimens with 2 to 9 teeth limited to upper one-quarter to one-half of anterior margin of maxilla; larg- er specimens with 14 to 20 teeth arranged along nearly entire anterior margin of max- illa. Largest cusp of teeth on maxilla re- curved ventrally. Scales cycloid, thin, relatively large. Lat- eral line distinctly decurved ventrally, com- pletely pored from supracleithrum to base of middle rays of caudal fin. Forty to 43 [42] scales in lateral line (70% of specimens with 42 scales); 6 or 7 [7] scales in transverse series from origin of dorsal fin to lateral line; 3 or 4 [3] scales in transverse series from origin of pelvic fin to lateral line (4 scales present in only 1 specimen); 3 or 4 [3] scales in transverse series from origin of anal fin to lateral line; 1 7 to 2 1 [20] scales along mid- dorsal line between tip of supraoccipital process and origin of dorsal fin (60% of spec- imens with 19 or 20 scales); 9 or 11 [11] scales along mid-dorsal line between pos- terior termination of dorsal fin and adipose fin (91% of specimens with 10 or 1 1 scales); 13 to 15 [13] horizontal scale rows around caudal peduncle (86% of specimens with 14 or 15 scales). Dorsal-fin rays ii,8 or 9 or iii,9 [ii,9] (ii,9 most common); anal-fin rays iv,3 1 to 35, or v,33 [iv,33] (iv,33 most common); pectoral- fin rays i,9 to 12 followed by to ii un- branched rays [i, 1 0,i] (i, 1 0,i most common); pelvic-fin rays i,6 [i,6]; principal caudal-fin rays 10/9 [10/9]. Dorsal fin profile obtusely acute, poste- rior unbranched and first branched ray sub- equal; posterior unbranched ray typically slightly longer. Dorsal fin situated on pos- terior half of body; origin of fin located slightly posterior of vertical line through an- terior terminus of anal fin, closer to base of caudal fin than to tip of snout. Longest di- mension of adipose fin approximately equal to horizontal width of pupil; origin of adi- pose fin slightly anterior of vertical line through posterior terminus of anal fin. Pec- toral fin large, profile distinctly acute; when fin depressed, tip extends to vertical line approximately two-thirds distance along pelvic fin. Pelvic fin profile obtusely acute, origin of fin at posterior margin of mid- ventral keel, tip of depressed fin extending posteriorly slightly beyond anus, but falling short of anterior terminus of anal fin. Ax- illary pelvic scale present, its length about one-third that of longest pelvic-fin ray. Cleared and counterstained 84.8 mm SL male with 6 to 15 basally directed bony hooks along posterior margins of first 5 branched pelvic-fin rays. Ventral margin of anal fin somewhat rounded anteriorly, with last unbranched and first branched rays 538 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON .'"tS^ Fig. 6. Moojenichthys myersi, Brazil, Bahia, Rio do Bra^o, 2 km SW of town of Rio do Brago, on Fazenda Luzia, USNM 304496, 33.4 mm SL. longest, subequal, following 10 to 12 branched rays rapidly decreasing in length, remaining anal-fin rays slowly decreasing in length. Males with 1 to 8 basally-directed bony hooks along posterior margins of dis- tal sections of 7 to 14 longest anal-fin rays. Caudal fin forked, lobes obtusely pointed. Total vertebrae 39 (2), 40 (20), 41 (2) [40]. Color in life. —Descriptions based on col- or transparencies of a series of recently pre- served specimens captured in August 1988 and February 1989. Overall coloration of specimens ranging from 15.9 to 33.4 mm SL clear to yellowish. Iris, lower jaw, in- fraorbital region, opercle, and peritoneum silvery. Tip of lower jaw, snout, dorsal por- tions of body and basal portions of caudal fin light yellow. Traces of yellow pigmen- tation apparent on dorsal, adipose, and anal fins. Other fins hyaline. Mid-lateral dark stripe on body quite obvious, but somewhat masked anteriorly by guanine. Dark stripe above anal fin, and dark pigmentation on all fins clearly visible. Specimens above approximately 50 mm SL bright silver overall, somewhat darker along dorsal portions of head and body. Dense guanine on scales completely mask- ing both dark stripes along mid-lateral sur- face of body, and those on ventrolateral sur- face of body above anal fin. Dark pigmentation on fins as in preserved spec- imens. Coloration in preservative. —Overall ground color of specimens fixed in formalin and lacking guanine on scales yellowish brown. Dense fields of small, dark chro- matophores on upper lip, snout, and dorsal surface of head (Figs. 5-7). Very intense, horizontally elongate stripe of dark pigmen- tation along dorsal portion of lower lip; less intense dark pigmentation ventral to this on lower jaw. Scattered small dark chromato- phores on lateral surface of head anterior to orbit and on opercle; pigmentation more obvious in specimens totally lacking gua- nine on head. Body with mid-lateral stripe of small, dark chromatophores extending from supra- cleithrum posteriorly to caudal peduncle; stripe gradually expanding vertically pos- teriorly; broadened into distinctly wider dif- fuse dark spot on lateral surface of caudal peduncle. Body dorsal of dark mid-lateral stripe with margins of scales outlined by dark chromatophores; pattern most ob- vious in medium sized specimens, some- what obscured by overall dusky appearance of dorsal portions of body in larger speci- mens. Stripe of dark chromatophores ex- tending posterodorsally along ventrolateral portion of body from slightly anterior of vertical through anterior terminus of fin posteriorly to posterior terminus of fin. Stripe wider and distinctly separated from base of anal fin anteriorly, becoming grad- ually narrower and approaching base of anal- fin rays posteriorly. VOLUME 103, NUMBER 3 539 n^ini i l i iip i ii R..-..^ Fig. 7. Moojenichthys myersi, Brazil, Bahia, Rio do Bra90, 2 km SW of town of Rio do Bra90, on Fazenda Luzia, MZUSP 40227, 81.6 mm SL. Dorsal fin with dense field of dark chro- matophores along distal portions of rays and membranes. Margins of adipose fin in larger specimens outlined by scattered dark chro- matophores. Anterior rays of pectoral and pelvic fins outlined by series of small dark chromatophores. Anterior margin and dis- tal portions of anal fin dusky, most rays outlined distally by dark chromatophores. Caudal fin dusky in specimens of all sizes, rays outlined by series of small dark chro- matophores. Common names.— Brazil, Bahia, Ilheus: "Mossarupe," "Piaba-faca," and "Cani- vete" (Miranda-Ribeiro, 1956:547). During the 1988 and 1989 expeditions the only name used by local fishermen was "Piaba- faca." £'C(9/o^.— Specimens collected during the 1988 and 1989 expeditions were collected in black waters containing limited suspend- ed material. The area surrounding the river was originally a portion of the Atlantic Coastal Forest, but much of the understory vegetation has been replaced by cocoa trees. In the areas sampled for fishes the Rio do Brago was between 1 and 25 m wide, ranged from 1.5 to 3 m deep, and had mats of floating vegetation along its margins. The bottom was sandy-mud with scattered boul- ders. Other fishes captured with Moojenichthys myersi and the families to which they are presently assigned were Steindachnerina elegans (Curimatidae); Nematocharax ve- nustus, Oligoscarcus macrolepis, Astyanax sp., Characidium. sp. (Characidae); Hoplias sp. (Erythrinidae); Rhamdia sp. (Pimelod- idae); Poecilia sp. (Poeciliidae); Astronotus ocellatus and Geophagus brasiliensis (Cich- lidae). The Astronotus ocellatus record rep- resents an introduction. Diet. — Examination of the stomach con- tents of the three cleared and counterstained specimens shows that the species eats mos- quito larvae and other aquatic inverte- brates. Distribution. —Known only from the Rio do Brago in the state of Bahia, Brazil. The original description of Moojenichthys my- ersi states that the type material was col- lected in the "Brago river, Ilheos." In ac- tuality the mouth of the "Brago river" [=Rio do Brago] is located approximately 7 km along the coast north of the city of Ilheus ("Ilheos" of Miranda-Ribeiro). Limited ichthyological collecting has taken place in the Bahian coastal drainages near the Rio do Bra90. Thus the lack of records o^ Moo- jenichthys myersi from other neighboring river systems may be a consequence of poor sampling. Material examined. —Brsizi]. Bahia: "Brago river," Ilheos [=Ilheus], MNRJ 540 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON 4127, 1 Specimen, holotype, 96.6 mm SL; same locality, MNRJ 4128, 1 specimen, paratype, 85.8 mm SL; Ilheus, Fazenda Pi- rataquice, MNRJ 5572, 1 specimen, 78.2 mm SL (locality not found in examined gaz- etteers or maps); Rio do Brago, 2 km SW of town of Rio do Bra90 (approx. 14°39'S, 39°16'W), on Fazenda Luzia, USNM 304497, 9 specimens, 20.5-84.8 mm SL (1 specimen, 84.8 mm SL, cleared and coun- terstained); USNM 304496, 15 specimens, 17.8-33.4 mm SL (1 specimen, 27.0 mm SL, cleared and counterstained); MZUSP 40226, 15 specimens, 15.9-30.5 mm SL; MZUSP 40227, 9 specimens, 19:6-84.1 mm SL (1 specimen, 56.0 mm SL, cleared and counterstained); ANSP 164287, 3 speci- mens, 17.4-26.5 mm SL; ANSP 164288, 2 specimens, 63.4-76.0 mm SL; MNRJ 11604, 3 specimens, 20.1-24.4 mm SL; MNRJ 1 1605, 2 specimens, 63.4-74.9 mm SL. Comparative cleared and stained material examined. —Brycon falcatus, USNM 226161, 2 specimens. Carnegiella strigata, USNM 225245, 5 specimens. Clupeacha- rax anchoveoides, USNM 302245, 1 spec- imen. Cynodon gibbus, USNM 270338, 2 specimens. Engraulisoma taeniatum, USNM 302225, 1 specimen. Gasteropelecus sternida, USNM 226337. Gnathocharax steindachneri, USNM 278995, 2 specimens. Rhaphiodon vulpinus, USNM 231549, 3 specimens. Triportheus angulatus, USNM 270343, 2 specimens. Triportheus sp., USNM 280498, 4 specimens; USNM 258079, 2 specimens. and Eduardo Castadelli, Jr. (all of FFCLRP- USP) who unstintingly assisted in the often difficult fishing efforts. Hertz F. Santos made a special second trip into the region of the type-locality to collect additional material of the species, including an extensive series that included most of the larger individuals reported on in this paper. Figures 5 to 7 were prepared by Mr. Theophilus B. Gris- wold. Sandra Raredon and Jeffi-ey Howe as- sisted in the research efforts at USNM. Ms. India Moreira (MNRJ) arranged for the loan of the holotype and paratype of Moojenich- thys myersi, and Mr. Luis Paulo S. Portugal (MZUSP) and Dr. Naercio A. Menezes (MZUSP) facilitated our examination of those specimens. This paper was improved by the suggestions of Dr. Stanley H. Weitz- man (USNM), Dr. Wayne C. Starnes (USNM), Dr. William L. Fink (University of Michigan, Museum of Zoology), Dr. Dar- rell J. Siebert (British Museum (Natural History)), and Mrs. Marilyn Weitzman (USNM). We thank all of the above for their assistance. Extensive logistical support for the ex- pedition that collected most of the speci- mens of Moojenichthys myersi was provid- ed by FFCLRP-USP. Funding for that collecting effort was provided through the Neotropical Lowland Research Program of the International Environmental Sciences Program of the Smithsonian Institution. That program also supported the research associated with this project both at mu- seums in Brazil and at the Smithsonian In- stitution. Acknowledgments A number of the specimens of Moojenich- thys myersi that served as the basis for this paper was collected during a collaborative FFCLRP-USP and USNM expedition in the state of Bahia, Brazil. The success of the collecting effort was assured by the enthu- siasm of Susan L. Jewett (USNM), Hertz F. Santos, Maura H. Manfrin, Eliseu B. Dias, Literature Cited Castro, R. M. C. 1981. Engraulisoma taeniatum, um novo genero e especie de Characidae da bacia do Rio Paraguai (Pisces, Ostariophysi).— Papeis Avulsos de Zoologia, Sao Paulo 34(1 1):135-139. Cope, E. D. 1872. On the fishes of the Ambiyacu River.— Proceedings of the Academy of Natural Sciences of Philadelphia (for 1871):257-269. Cuvier, G. & A. Valenciennes. 1 849. Histoire natu- relle des poissons. Volume 22:1-395. VOLUME 103, NUMBER 3 541 Fink, W. L., & S. H. Weitzman. 1974. The so-called cheirodontin fishes of Central America with de- scriptions of two new species (Pisces: Charac- idae). — Smithsonian Contributions to Zoology 172:1^6. Fowler, H. W. 1 906. Further knowledge of some het- erognathous fishes. Part II.— Proceedings of the Academy of Natural Sciences of Philadelphia (for 1906):43 1-^88. Gery, J. 1972. Poissons characoides des Guyanes, I. Generalitees. II. Famile des Serrasalmidae. — Zoologische Verhandelingen 122:1-250. . 1977. Characoids of the world. TFH Publi- cations, Neptune City, New Jersey, 672 pp. Greenwood, P. H., D. E. Rosen, S. H. Weitzman, & G. S. Myers. 1966. Phyletic studies of teleos- tean fishes, with a provisional classification of living forms. — Bulletin of the American Mu- seum of Natural History 131(4):339^56. Howes, G. J. 1976. The cranial musculature and tax- onomy of characoid fishes of the tribes Cyno- dontini and Characini. — Bulletin of the British Museum (Natural History), Zoology 29:203-248. . 1982. Review of the genus 5r>'co«(Teloestei: Characoidei). — Bulletin of the British Museum (Natural History), Zoology 43(1): 1^7. Miquelarena, A. H., & J. R. Casciotta. 1982. Presen- cia en la Argentina de Clupeacharax anchov- eoides Pearson, 1924. — Limnobios 2(5):333- 335. Miranda-Ribeiro, P. de. 1941. Notas para o estudo dos characinidos brasileiros (Peixes— Genero Chalcinus Cuv. & Val. 1849). — Papeis Avulsos do Departamento do Zoologia 1(18):159-174. . 1956. On a new genus and new species of South American fishes.— Proceedings of the 14th International Congress of Zoology, Copenha- gen, 16(Vertebrates):546-547. Myers, G. S. 1 940. Suppression ofAcropsis and Chal- cinus, two preoccupied generic names of South American fresh-water fishes.— Stanford Ichthy- ological Bulletin 1(5): 170. Ortega, H., & R. P. Vari. 1986. Annotated checklist of the freshwater fishes of Peru.— Smithsonian Contributions to Zoology 437:1-25. Pearson, N. E. 1924. The fishes of the eastern slope of the Andes. I. The fishes of the Rio Beni basin, Bolivia, collected by the Mulford Expedition.— Indiana University Studies 1 1(64): 1-63. Poll, M. 1957. Les genres des poissons d'eau douce de I'Afrique.— Annales du Musee Royal du Congo Beige, Series Zoologiques 54:1-191. Regan, C. T. 1911. The classification of the teleostean fishes of the order Ostariophysi.— 1. Cyprinoi- dea.— Annals and Magazine of Natural History, ser. 8, 8:14-32. Roberts, T. R. 1969. Osteology and relationships of the characoid fishes, particularly the genera Hepsetus, Salminus, Hoplias, Ctenolucius, and Acestrorhynchus.—VrocQcdmgs of the Califor- nia Academy of Sciences, fourth series, 36(15): 391-500. . 1973. Osteology and relationships of the Pro- chilodontidae, a South American family of cha- racoid fishes. — Bulletin of the Museum of Com- parative Zoology 145(4):213-235. . 1974. Osteology and classification of the neo- tropical characoid fishes of the families Hemi- odontidae (including Anodontinae) and Paro- dontidae. — Bulletin of the Museum of Comparative Zoology 146(9):41 1^72. Sazima, I. 1983. Scale-eating in characoids and other fishes.— Environmental Biology of Fishes 9(2): 87-101. Schultz, L. P. 1 944. The fishes of the family Chara- cinidae from Venezuela, with descriptions of seventeen new forms. — Proceedings of the United States National Museum 95:235-367. Starks, E. C. 1930. The primary shoulder girdle of the bony fishes. — Stanford University Publica- tions, University Series, Biological Sciences 6(2): 149-239. Stewart, D. J., R. Barriga S., & M. Ibarra. 1987. Ic- tiofauna de la cuenca del Rio Napo, Ecuador Oriental: lista anotada de especies. — Politecni- ca, revista de Informacion Tecnico-Cientifica, 12(4):9-63. Vari, R. P. 1977. Notes on the characoid subfamily Iguanodectinae, with a description of a new species.— American Museum Novitates 2612: 1-6. . 1979. Anatomy, relationships and classifi- cation of the families Citharinidae and Distich- odontidae (Pisces, Characoidea). — Bulletin of the British Museum (Natural History) 36(5):261- 344. . 1986. Serrabrycon magoi, a new genus and species of scale-eating characid (Pisces: Characi- formes) from the upper Rio Negro. — Proceed- ings of the Biological Society of Washington 99(2):328-334. Weitzman, S. H. 1954. The osteology and relation- ships of the South American characid fishes of the subfamily Gasteropelecinae. — Stanford Ich- thyological Bulletin 4(4):2I3-263. . 1960. The phylogenetic relationships of Tn^- ortheus, a genus of South American characid fishes. — Stanford Ichthyological Bulletin 7(4): 239-244. . 1962. The osteology of Brycon meeki. a gen- eralized characid fish, with an osteological def- inition of the family.— Stanford Ichthyological Bulletin 8(1): 1-77. . 1964. Osteology and relationships of the South 542 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON America characid fishes of the subfamihes Le- biasininae and Erythrininae with special refer- ence to the subtribe Nannostomina.— Proceed- ings of the United States National Museum 116: 127-170. - S. V. Fink, & N. A. Menezes. 1985. Appendix 3. A critique of the monophyly of the Glandu- locaudinae. Pp. 112-121 inS. H. Weitzman and S. V. Fink, eds., Xenurobryconin phylogeny and putative pheromone pumps in glandulocaudine fishes (Teleostei: Characidae). Smithsonian Contributions to Zoology 421:1-121. -, & W. L. Fink. 1983. Relationships of the neon tetras, a group of South American fresh- water fishes (Teleostei, Characidae), with com- ments on the phylogeny of New World characi- forms.— Bulletin of the Museum of Comparative Zoology 150(6):339-395. , & R. P. Vari. 1987. Two new species and a new genus of miniature characid fishes (Teleos- tei: Characiformes) from northern South Amer- ica.— Proceedings of the Biological Society of Washington 100:640-652. (RMCC) Departmento de Biologia, Fa- culdade de Filosofia, Ciencias e Letras de Ribeirao Preto, Universidade de Sao Paulo, Avenida Bandeirantes 3900, 14049 Ribei- rao Preto SP, Brazil; and (RPV) Department of Vertebrate Zoology (Fishes), National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560.