Austrolebias Costa

Wilson J. E. M. Costa, 2006, The South American annual killifish genus Austrolebias (Teleostei: Cyprinodontiformes: Rivulidae): phylogenetic relationships, descriptive morphology and taxonomic revision., Zootaxa (1213), pp. 1-162 : 6-9

[[ Genus Austrolebias Costa ]]

The Neotropical aplocheiloid genus Austrolebias Costa (including Megalebias Costa as herein proposed) comprises a diversified clade of annual killifishes, inhabiting seasonal pools formed during rainy seasons in southern Brazil, Paraguay, Uruguay and northern and northeastern Argentina (Costa, 1998a, 2002a). In contrast to other aplocheiloids that are geographically restricted to warm tropical areas of the world, most species of Austrolebias are endemic to temperate areas of South America, where rainy seasons are concentrated in winter months. Consequently, they are found in cold waters, sometimes near the freezing point.

Because of the bright blue colors of some species and their resistance to low temperatures, some species of Austrolebias have become popular aquarium fishes in Europe since the early 20th century (e. g., Myers, 1952). Some species are also interesting because of their possession of morphological features uncommon, or absent from, other aplocheiloids, such as a membranous attachment between male urogenital papilla and anal fin, medial fusion of pelvic fins, presence of prominent contact organs on scales and unpaired fins, and loss of scales on head and trunk (e. g., Amato, 1986; Costa, 2002a). More remarkable is the large size and development of the jaws in some species of the subgenus Megalebias , making them voracious predators of closely related species, an uncommon feature among aplocheiloids.

All species of Austrolebias are annual fishes, living in temporary pools and swamps formed during the rainy season. In the Pampas region, including Uruguay, Buenos Aires and Entre Ríos provinces of Argentina and southern Brazil, pools are usually formed during the winter months (June -July), whereas in the Chaco region of northern Argentina and Paraguay, pools are mainly formed during the summer months (December -January). Individuals become adults about two months after formation of the pools, and die when the pools dry, usually during late spring in the Pampas and during the winter in the Chaco. Chorion-thickened eggs survive during the dry season and hatch just after beginning the rainy season.

All species were formerly placed in Cynolebias prior to descriptions of the genera Austrolebias and Megalebias by Costa (1998a). As a result of their early popularization as aquarium fish, several nominal species were erected during the 20th century that were poorly described in the aquarium literature, with the most important diagnostic features completely omitted (see History below). Because of these poor descriptions and usually inaccessible study material, the taxonomy of Austrolebias is probably the most problematic of all Neotropical annual fishes. In addition, aquarium papers based on recent explorations of annual fish habitats suggest the existence of numerous undescribed species (e. g., Reichert, 1994a; Reichert et al., 1997). The objectives of the present study are: to provide a description of morphological traits in Austrolebias ; to test previous phylogenetic hypotheses on the basis of reexamination of morphological characters; and to provide a taxonomic revision of the genus based on large recent collections.

Taxonomic history

In contrast to all other South American areas, from where most annual fish species were discovered and described after 1940, annual fish habitats in Buenos Aires province (Argentina) were frequently sampled during the late 19th and early 20th centuries. Steindachner (1881) first described species today placed in Austrolebias : Cynolebias elongatus , C. bellottii , and C. maculatus , all from Buenos Aires province.

Subsequently, Günther (1883) described C. robustus , also from Buenos Aires province. This species was compared with C. porosus , described by Steindachner (1876), but no mention was made of other species from the same area, described by the same author two years before.

Garman’s (1895) revision of cyprinodontiform fishes was not relevant to taxonomic knowledge of the cynolebiatins. By contrast, Garman equivocally placed C. elongatus in the synonymy of C. porosus Steindachner (a species endemic to northeastern Brazil), and placed C. robustus in the synonymy of C. maculatus .

Berg (1897), in a broad systematic account on South American fishes, was the first to report the extreme sexual dimorphism characteristic of cynolebiatins. He then noted that C. maculatus , based on females, was in fact a synonym of C. bellottii , which was based on males. In the same paper, two other species from Buenos Aires province, C. gibberosus and C. holmbergi , were described. Berg followed Garman (1895) in considering C. robustus a synonym of C. bellottii , but recognized C. elongatus as a valid species.

Regan (1912) revised the Neotropical killifishes today placed in the Rivulidae. New data were provided for some known Argentinean taxa, and a new species, C. nigripinnis , from Buenos Aires province was described. All nominal species except C. maculatus were considered valid.

During the 1920’s, Alfred Adloff explored annual fish habitats in southern Brazil, around the city of Porto Alegre, where he discovered two species, which were sent to Germany and described by Ahl (1922, 1924) as C. adloffi and C. wolterstorffi . Material from southern Brazil and Argentina comprised the basis for Ahl’s revisionary studies on Cynolebias (Ahl, 1922, 1934). However, these studies were based on limited morphological data, making it difficult to recognize species, and the validity of most taxa was dubious. Ahl (1934) also described another species from the Buenos Aires area, C. spinifer , and a species with possibly equivocal type locality, C. schreitmuelleri , supposedly from Rio de Janeiro, southeastern Brazil. Ahl (1938) described C. irregularis from material probably collected in Argentina, further increasing taxonomic problems in the group.

During the 1940’s, the known geographic range of the group was significantly increased with the discovery of C. carvalhoi in the highlands of the rio Iguaçu valley, at about 800m altitude, in southern Brazil, by George Myers and Antenor de Carvalho (Myers, 1952). Unfortunately, this species was obscurely described, having become available only by its brief mention in a paper on Amazonian fishes published in an aquarium magazine (Myers, 1947).

During the 1960’s and 1970’s, several species were reported from Uruguay and Argentina, including some interesting new species, beginning an important productive period for the taxonomic knowledge on the group. Raúl Vaz-Ferreira and collaborators provided some noteworthy taxonomic contributions, including the descriptions of C. viarius , C. cheradophilus and C. luteoflammulatus from the coastal plains of eastern Uruguay (Vaz-Ferreira et al., 1964), and several new records for Uruguay (Vaz-Ferreira and Sierra, 1971). Vaz-Ferreira and Sierra (1973) considered C. holmbergi to be a synonym of C. elongatus , and C. gibberosus and C. irregularis as synonyms of C. bellottii . Ringuelet et al. (1967) recorded cynolebiatins from several Argentine localities, including northern regions in the Chaco. Taberner et al. (1974) described C. nonoiuliensis , and Castello and Lopez (1974) described C. alexandri , both from northeastern Argentina.

After 1980, South American annual fish habitats were intensively sampled, making available new data for several taxonomic publications. Ten new species were described from Uruguay by Amato (1986) ( C. gymnoventris , C. melanoorus , C. affinis , C. prognathus , and C. cinereus ), Berkenkamp et al. (1994, 1997) ( C. vazferreirai and C. nioni ), Loureiro & García (2004) ( C. reicherti ), Loureiro et al. (2004) ( A. arachan ) and Perujo et al. (2005) ( A. luzardoi ). Three were described from Paraguay by Huber (1995) ( C. patriciae , C. vandenbergi , and C. monstrosus ); and eight were described from southern Brazil by Amato (1987) ( C. cyaneus ), Costa, (1999a, b) ( C. ibicuiensis and C. periodicus ), Costa and Cheffe (2001, 2002) ( A. charrua , A. minuano , A. nigrofasciatus , and A. jaegari ); and Costa et al. (2004) ( A. varzea ). Some species were also redescribed after new collections (Costa, 1998b, 2002b; Costa and Cheffe, 2001, 2002). However, recent advances do not include clarification of older taxonomic problems. Validity of old names, such as C. robustus and C. holmbergi , has been the focus of some debate in the aquarium literature (e. g., Lazara, 1981; Wildekamp, 1995); this has sometimes been based on equivocal assumptions, which have further confused certain taxonomic issues (see remarks under descriptions of A. robustus , A. nonoiuliensis , A. wolterstorffi , and A. elongatus below). In addition, a series of aquarium papers based on intensive collections directed to annual fish habitat in Uruguay reports great color pattern variability among different populations, which suggests the existence of several undescribed species (Reichert, 1994a, b; Reichert et al., 1997).

Species now placed in Austrolebias have been divided into different subgroups. The larger species have usually been considered a distinct subgroup (e. g., elongatus group in Amato, 1986), more closely related to species of Cynolebias than to the smaller Austrolebias species (e. g., Vaz-Ferreira et al., 1964; Amato, 1986; Costa, 1995, 1998a; Wildekamp, 1995). Among the smaller species, Amato (1986) recognized two species groups (the adloffi and the luteoflammulatus groups), defined by a combination of characters of body depth, number of dorsal and anal-fin rays, and presence of contact organs, and a third group (not named) with intermediate features.

Recent phylogenetic analyses at the species level, based on morphological characters (Costa, 2002a) and mitochondrial genes ( García et al., 2000, 2002) corroborate the monophyly of Austrolebias as herein proposed. However, species clades proposed in each paper are highly conflicting in most aspects (see discussion below for a comparison among hypotheses).