Zootaxa, Pterolebias
Total Page:16
File Type:pdf, Size:1020Kb
Zootaxa 1067: 1–36 (2005) ISSN 1175-5326 (print edition) www.mapress.com/zootaxa/ ZOOTAXA 1067 Copyright © 2005 Magnolia Press ISSN 1175-5334 (online edition) The Neotropical annual killifish genus Pterolebias Garman (Teleostei: Cyprinodontiformes: Rivulidae): phylogenetic relationships, descriptive morphology, and taxonomic revision WILSON J. E. M. COSTA * * Laboratório de Ictiologia Geral e Aplicada, Departamento de Zoologia, Universidade Federal do Rio de Janeiro, Caixa Postal 68049, CEP 21944-970, Rio de Janeiro, Brasil. [email protected] Abstract Pterolebias is revised on the basis of osteology, external morphology, latero-sensory system, and color patterns. Two species are recognized as valid: P. longipinnis, from the eastern and southern Amazon River basin and the ParanáParaguay River system, in Brazil, Bolivia, Paraguay and Argentina; and P. phasianus, from the Paraguay River basin, in Brazil and Bolivia. Pterolebias bokermanni and P. luelingi are considered synonyms of P. longipinnis. Monophyly of Pterolebias is corroborated by the morphology of the angulo-articular, second pharyngobranchial, maxilla, metapterygoid, quadrate, basihyal, and two derived color patterns. In contrast to molecular studies, monophyly of an assemblage including Pterolebias and Gnatholebias is herein strongly supported by reduction of the interarcual cartilage, morphology of the anterior proximal radials of the anal fin and pelvic girdle, presence of scale rows on the anal-fin base, derived jaw dentition, numerous pelvic-fin rays, and long pelvic fins. A recent proposal to unite Pterolebias and Gnatholebias in a single genus is rejected. Key words: Killifishes, Cyprinodontiformes, Rivulidae, Pterolebias, Neotropical, Amazon, Pantanal, systematics, phylogenetic relationships Resumo Pterolebias é revisado com base em osteologia, morfologia externa, sistema látero-sensorial e padrões de colorido. Duas espécies são reconhecidas como válidas: P. longipinnis, do leste e sul da bacia amazônica e do sistema fluvial ParanáParaguai, no Brasil, Bolívia, Paraguai e Argentina, e P. phasianus, da bacia do rio Paraguai, no Brasil e Bolívia; P. bokermanni e P. luelingi são consideradas sinônimos de P. longipinnis. Monofiletismo de Pterolebias é corroborado pela morfologia do ângulo-articular, segundo faringo-branquial, maxila, metapterigóide, quadrado, basial e dois padrões de colorido derivados. Ao contrário de estudos moleculares, monofiletismo de um agrupamento incluindo Pterolebias e Gnatholebias é aqui altamente sustentado pela redução de cartilagem interarcual, morfologia dos radiais proximais anteriores da nadadeira anal e da Accepted by C. Gilbert: 16 Sept. 2005; published: 21 Oct. 2005 1 ZOOTAXA cintura pélvica, presença de escamas na base da nadadeira anal, dentição derivada em maxilas, 1067 numerosos raios na nadadeira pélvica e longa nadadeira pélvica. Uma recente proposta para unir Pterolebias e Gnatholebias num único gênero é rejeitada. Introduction Pterolebias Garman, 1895, occurs over a vast area of South America, including the eastern and southern Amazon River basin and the ParanáParaguay River system (Costa, 1998a). It is a well defined clade of annual killifishes, supported both by morphological (Costa, 1998a) and molecular studies (Murphy et al., 1999; Hrbek & Larson, 1999). However, its taxonomy is still poorly known. It includes four nominal species: P. longipinnis Garman, 1895; P. bokermanni Travassos, 1969; P. luelingi (Meinken, 1969); and P. phasianus Costa, 1988. Pterolebias longipinnis has been recorded over the entire geographic range of the genus, from Marajó Island in northern Brazil (about 1º S) to Corrientes in Argentina (about 28° S). Two nominal species from the Madeira River basin, P. bokermanni and P. luelingi, are insufficiently diagnosed and not distinguishable from P. longipinnis (Thomerson, 1984; Costa, 1988). Pterolebias phasianus is a distinctively slender species with a unique color pattern. Relationships of Pterolebias to other rivulid genera are also controversial. The hypothesis generated by morphological phylogenetic analysis indicates that Pterolebias is the sister group to Gnatholebias Costa, 1998, an annual killifish genus occurring in the Orinoco River basin and adjacent coastal river basins in Venezuela and Colombia (Costa, 1998a). However, hypotheses derived from molecular studies support Gnatholebias as more closely related to other annual rivulid genera endemic to northern South America (i.e., Renova Thomerson & Taphorn, 1995; Terranatos Taphorn & Thomerson, 1978; Micromoema Costa, 1998; Rachovia Myers, 1927; and Austrofundulus Myers, 1932) than to Pterolebias (Murphy et al., 1999; Hrbek & Larson, 1999). The objectives of the present study are: 1) to provide a detailed description of morphological traits of Pterolebias, which in turn allows better morphological character analyses; 2) to identify informative characters useful for the diagnoses of species, based on large collections totaling over 400 specimens over the entire geographic range of the genus; and 3) to recognize and redescribe the valid species. Taxonomic History Pterolebias was first described to include a single species, P. longipinnis Garman, collected in the eastern Brazilian Amazon during the Thayer Expedition (18651866) (Garman, 1895). In Garman’s key to identification of cyprinodontiform genera, Pterolebias was distinguished from the other two rivulid genera known at that time, Rivulus Poey, 1860, and Cynolebias Steindachner, 1876, by having “body sharp-edged 2 © 2005 Magnolia Press COSTA behind vent”. This condition was repeatedly listed as diagnostic for Pterolebias (e. g., ZOOTAXA Regan, 1912; Myers, 1927), but it was considered to be a possible artifact of preservation 1067 by Myers (in Weitzman & Wourms, 1967). In the 20th century, nine new species over a wide area of South America, including the Orinoco, Amazonas and Paraguay river basins, were assigned to the genus Pterolebias: P. peruensis Myers, P. wischmanni Seegers, and P. rubrocaudatus Seegers, from the Peruvian Amazon (Myers, 1954; Seegers, 1983, 1984), P. bokermanni from the Madeira River drainage of Brazil (Travassos, 1955), P. zonatus Myers, P. maculipinnis Radda, and P. hoignei Thomerson from the Orinoco basin of Venezuela (Myers, 1935; Radda, 1964; Thomerson, 1974), P. staecki Seegers from the Brazilian central Amazon basin (Seegers, 1987), P. phasianus Costa from the Paraguay River basin, Brazil (Costa, 1988), and P. obliquus Costa, Sarmiento & Barrera from the Bolivian Amazon (Costa et al., 1996). When new collections of P. longipinnis were made in regions distant from the type locality, including the Paraná-Paraguay River basin of Argentina and Paraguay (Alonso- de-Aramburu, 1961), this species became the most geographically widespread South American annual fish known. During the 1970s, a new annual fish species of unknown South American origin was reported in the aquarium literature as Pterolebias NSC-1 (e. g., Terceira, 1973). Costa (1989) recognized it as an undescribed genus and species, which he described as Moema piriana Costa from the eastern Brazilian Amazon. Thomerson (1984), following comparison of the original description of Rivulichthys luelingi Meinken, from the Chapare River floodplains, Madeira River drainage, Bolivian Amazon, with data obtained from the examination of type specimens of P. longipinnis, concluded that the former is a synonym of the latter. Costa (1988) examined the type specimens of P. bokermanni and did not find characters useful for distinguishing it from P. longipinnis, but preferred not to synonymize these species, since data on live color patterns were not available at that time. However, Huber (1995) identified Pterolebias sp. aff. longipinnis from the Paraguay River basin, thus suggesting it to be a distinct species, but he provided no justification for his opinion. Schindler & Staeck (1993a, b) provided data on recent collections of Pterolebias, and, in addition, Schindler (2004) recognized Pterolebias luelingi as a valid species, and reported diagnostic features for P. bokermanni and P. longipinnis. For almost 60 years after Garmans paper, no additional diagnostic data were provided for the genus Pterolebias and no hypothesis of relationships was erected. Myers (1954) suggested Pterolebias as being closely related to Rachovia, due to the common occurrence of a short pelvic fin in Rachovia and in P. peruensis. However, a short pelvic fin is the primitive condition for rivulids, occurring in most aplocheiloid taxa, and does not, by itself, indicate a close relationship between Pterolebias and Rachovia. Weitzman & Wourms (1967) hypothesized that Rachovia, Pterolebias and Austrofundulus (including Terranatos) form a natural assemblage, based on their common PTEROLEBIAS © 2005 Magnolia Press 3 ZOOTAXA geographic distribution in northern South America and presence of scales on the caudal 1067 fin. However, although both Rachovia and Austrofundulus have scales extensively covering the caudal fin, this condition does not occur in Pterolebias, but is present in other rivulid genera (i. e., Neofundulus Myers and Trigonectes Myers); therefore, it does not support monophyly of the group including Rachovia, Pterolebias and Austrofundulus. Weitzman & Wourms (1967) also proposed a group including Cynolebias, Cynopoecilus Regan, Pterolebias, Rachovia, Austrofundulus, Leptolebias Myers, and Simpsonichthys de Carvalho, based on the possession of thickened rays on the anterior half of the anal fin in females. According to Parenti (1981),